Contents Introduction Part 1: 150 most important positions in the Opening and the Middlegame 1-30: Development of the Pieces 31-64: Exchange of Material 65-108: Manoeuvring 65-78: The Knight Manoeuvre 79-87: The Bishop Manoeuvre 88-98: The Rook Manoeuvre 99-102: The Queen Manoeuvre 103-108: Manoeuvring with Several Pieces 109-130: Pawn-Play and the Centre 109-114: Pawn-Play on the Wing 115-130: Pawn-Play in the Centre 131-147: Play on the Wings 131-135: The Minority Attack 136-137: Queenside Attack 138-147: Kingside Attack 148-150: Psychological and Pragmatic Moves Part 2: 150 most important positions in the Ending 151-167: Pawn Endings 168-180: Knight Endings 181-191: Bishop Endings 192-205: Knight and Bishop Endings 206-277: Rook Endings 278-300: Queen Endings Introduction The 300 most important chess positions, demonstrating positional ideas in the opening, middlegame and ending, was originally an email course which I’ve run on two occasions. The course was labelled “Less is more” and was aimed at Swedish chessplayers, especially the juniors at my club SK Rockaden. The main idea was to study five positions very carefully each week and then meet on a regular basis and discuss some of them, which would sometimes be the starting point for training games by my students. During the preparation of this book all the positions were checked for accuracy, after which some were omitted and replaced whilst still retaining more than two thirds from the original material. However, one important change in the present book is that there are now 150 positions representing the endgame and the same number for the opening and the middlegame combined. Originally the quota was 50 openings, 150 middlegames and 100 endgames. The reason I made the changes was twofold: firstly, I thought too many important endgame positions were missing and, secondly, it’s hard to explain what is an opening position and a middlegame position and draw a consistent line of demarcation. According to Alexei Suetin the middlegame begins when an attack is launched and I’m inclined to agree. This means, for example, 1 e4 c5 2 ♘f3 d6 3 d4 cxd4 4 ♘xd4 ♘f6 5 ♘c3 g6 6 ♗e3 ♗g7 7 f3 0-0 8 ♕d2 ♘c6 9 ♗c4 ♗d7 10 0-0-0 ♖c8 11 ♗b3 ♘e5 12 h4 is a middlegame position because the targets are already clear for both players. This also implies that the opening moves 1 e4 e5 2 ♘f3 ♘c6 3 ♗c4 ♘f6 4 ♘g5 represent a middlegame for White but an opening for Black. Such reasoning might well be confusing for a less advanced player so I have chosen to group openings and middlegames together without defining what the specific phase really is. For example, rapid piece development is more important in open positions but in closed positions the manoeuvre may be more important, thereby giving the opening more of a middlegame character. The hypermoderns even regarded the opening as a middlegame and since at a higher level it’s not clear what is what, I think it was most appropriate to treat the two phases as one. Which phase is the most important one to study? In my opinion it’s the endgame because in that phase it’s easier to see how the chess pieces cooperate. By learning back to front, so to speak, it’s easier to understand how the pieces work best and how they should harmonise with each other in the opening or the middlegame – unless the balance has been seriously disturbed, when some kind of attack is the order of the day. If you really want to learn how to play with bishop and knight there is no better way than to study the three different methods of checkmating the king. If you want to learn how to play with the bishop pair you should study the rare but important endgame two bishops versus one knight. Another reason for stressing the value of the endgame is that multipiece endings need greater coverage since this type of ending is more or less ignored in chess literature. All endings with two pieces on each side are to a large extent under-represented. The same goes for three and four pieces each as well. According to Speelman’s clever definition of the endgame, the total value of the pieces involved should not be more than thirteen points. This means that four minor pieces is regarded as an endgame but, strictly speaking, queen and rook is not – because of the increased danger to the king. As is well-known, the activity and safety of the king are other criteria to be taken into account in the endgame. In the present work, despite Speelman’s definition, I have chosen to regard all positions with two pieces each as endgames even if major pieces are involved and the king is far from safe. The first time I became aware of the whole concept of positions containing positional ideas was when I studied Training for the Tournament Player by Dvoretsky and Yusupov in the 90s. The authors illustrated this highly important concept with a few positions. At the time I wasn’t fully aware of the importance and high practical value of the concept even though I had to a certain extent practised it unconsciously as a youngster. However, as time went by I realised that this has to a large extent been a forgotten and underestimated way of speeding up progress in chess. Surprisingly, even to this day there is still no book focusing on this interesting concept of positional sketches. Perhaps one reason is that it’s been a well-hidden secret among the best trainers. There are a few books by Lev Alburt with the 300 most important positions or 320 key positions but these mainly focus on tactics and are more fundamental in their approach. One of his books is actually titled 300 Most Important Positions and Ideas. In addition there is another book by Ziyatdinov with the title GMRam: Essential Grandmaster Chess Knowledge, which presents 300 positions, but unfortunately there are no solutions to them. Not only is a knowledge of the most relevant positions and ideas in itself important but also the solutions should be presented in a pedagogical and easy-to-learn fashion in order to facilitate comprehension. After all, you are supposed to learn, understand and remember the positional ideas for all time and therefore you need to integrate the positions into your conscious chess thinking as early as possible in your life. Of course if you’ve been unable to learn appropriate chess thinking as a youngster then it’s still better to learn the ideas later, regardless of your age, than never to learn them at all. To benefit from this approach to learning it’s vital to repeat the positions from time to time by looking at them again to see if you can instantly remember the ideas. If you have forgotten the solution or the technique required you should check it again and again until these are permanently embedded in your mind. Positional ideas are important in all phases of the game and that obviously includes the transitional phases going from the opening to the middlegame and from the middlegame (or multi-piece endgame) to the endgame. Other related material can be found in two recent books by Arthur Van de Oudeweetering: Train Your Chess Pattern Recognition – Key Moves and Motifs in the Middlegame. These focus on the opening and the middlegame and concentrate on chess pattern recognition which is another important component in the present book as well. However, here I want to distill the most important positional ideas and techniques and focus solely on these. In other words, to facilitate chess training and make it more effective and concise. Why study so many positions if a smaller number is sufficient? By learning one key position, one idea or one technique you can apply it in many other situations. The reason for the limited number of only 300 positions is motivated by the phrase “less is more”. It’s also in accordance with the principle that the less you know the less you’ll forget. In other words it will be easier to remember 300 positional ideas rather than, let’s say, 1,000 ideas since you’ll have a smaller number of positions to learn and repeat on a regular basis. I was first introduced to this clever training concept as a junior at the end of the 70s. My club SK-33 was located in the small town of Enköping and it helped me and a friend of mine to arrange a couple of meetings with Robert Danielsson, at the time a wellknown specialist in teaching chess to children in Stockholm. He had written books and study material for beginners on a high pedagogical level. He was the right person to approach for chess hungry players like myself and Peter Fransson, who by the way later won the Swedish Junior Championship in 1980. When we met we agreed to follow a strict and structured system which necessitated the study of five positions from Pachman’s Mittelspielpraxis im Schach and Endspielpraxis im Schach very carefully every week. If you follow such a schedule ambitiously it will take you approximately one year to assimilate the knowledge to be gained from the positions and utilise all this for a lifetime. The reason for studying only five positions was to focus on those that are the most important and to concentrate solely on them. If you study too many positions and gloss over them too quickly you run an increased risk of eventually losing your discipline and then forget what you once endeavoured to learn. The key is to learn slowly to give you time to reflect on the positions and if you can also discuss them with your friends so much the better. The best way to learn is actually to teach others so if you have this opportunity you should take advantage of it. Another way of learning, particularly endgame positions, is to play them out with a friend or a computer from both sides of the chessboard. How do we know what are the 300 most important positions (positional ideas) in chess? Of course this is a relative and dynamic notion depending on your previous knowledge and experience. For a novice, a position showing how to mate with a queen against a king is important, but if you already know how to do this then it becomes superfluous and even a waste of time, unless you really do feel you need to review the process. This book is aimed at a rating category from 1000 to 3000, which means I have chosen to satisfy all levels at the same time. This also implies that the lower the individual level, the more positions there are to be studied. At the highest level, around 3000 Elo rating, it means you must be able to beat the computer when you have a queen versus a rook or when you are in possession of two bishops versus a knight. You must be able to defend in rook endings where you are two pawns down (rook pawn and bishop pawn) in time pressure, as well as holding the draw in queen endings where you are one or two pawns down, especially the notorious endgame rook pawn and knight pawn. One way of addressing different levels at the same time is to start with an advanced position and then, as things become clearer, wait for something more basic to turn up. In such a situation there might be two numbered diagrams for the same game or they might be part of the variations at the end of the game. Among the 300 most important positions you will not find any diagram or variation on how to win with a queen versus a lonely king. Nor will you find a numbered diagram on how to win with a queen against a knight. However, you will find one important position showing how to play with a queen against two knights and in the variations you will be shown how to win against one knight. The only difference compared with a lone king is that you don’t need to think about pitfalls such as stalemate. You simply ignore the knight and play for mate rather than to win the knight. So, by showing a more advanced position, some easier ones will appear in the variations. The following position is not regarded as important but it is embedded in the variations of the more complicated version where Black has two knights and therefore that position is regarded as the more important one. Tablebase The technique when playing against a lone knight is the same as if the king were alone, but in a way it’s made easier since there is no stalemate trap. The trick is to play for mate rather than to catch the knight: 29 ♔f6 ♔e3 30 ♕e8+ ♔f3 31 ♕e7 ♘e2 32 ♕e5 ♘g1 33 ♕d4 ♘e2 34 ♕d3+ ♔f2 35 ♔f5 ♘g1 36 ♔e4 ♔g2 37 ♕d1 ♘h3 38 ♕f3+ ♔h2 39 ♔e3 ♘f4 40 ♕f2+ ♔h3 41 ♔f3 ♘h5 42 ♕g1 ♘f6 43 ♕g3 mate. Now you might perhaps argue that it’s very rare, if ever, to get the kind of material distribution where you have a queen against one or two knights, and that’s certainly true. I have never had it myself in a serious game but this fact is irrelevant when it comes to deciding whether a position is important or not. The whole concept is built on the idea that it isn’t only a matter of whether you encounter this or that position in practical play. What is of more and even vital importance is to gain a deeper appreciation of the inner qualities of the pieces, their movements and their actual value in any specific position. For example, the following famous study by Bilguer is presented as one of the 300 most important positions where you are supposed to understand, grasp and assimilate the most important idea(s). The queen and king are fighting against two knights and a king, and the only way to win and the important idea to grasp is to cut off the king on the last rank. Bilguer 1843 White to move According to Pachman and Chéron White should enter the heart of Black’s position and play the centralising 1 ♕e6? but this actually leads only to a draw after 1...♔g7. However, it’s a forced mate in 43 moves if White plays the logical move 1 ♕c7! cutting off the king on the last rank. The only way for the defending king to advance is by shielding it from the queen’s control of the seventh rank. which would mean placing the knights in passive positions. Note that White has reduced the value of the queen to the function of a mere rook. The centralising queen move to e6 is actually a mistake since after this Black can coordinate his forces. The moral here is that the queen should not act like a queen in this position. It should be humble and just behave like a rook, only controlling the seventh rank. The queen doesn’t need to demonstrate any of its diagonal powers yet. From time to time, even great players overlook the importance of the seventh rank when a queen is involved. Sometimes we have to forget that a queen is a queen and devalue it, transforming it into a mere rook or, can you imagine, a bishop. A truly strong queen exploits the fact that it can change roles between all the pieces of the board except the knight. Such a queen is for example the Rétiqueen where the queen acts like a bishop in the corner of the board. Here is one instructive example where Morphy, as a twelve-yearold, failed to use the queen as a rook on the seventh rank. P. Morphy – A. Morphy New Orleans 1949 White to move Morphy played the nice and quiet centralising move 19 ♕d5. He missed the faster and prettier king hunt 19 ♖xf7+ ♔xf7 20 ♕d7+ (Note that the queen is like a rook on the seventh.) 20...♔f6 (The variation 20...♔g8 21 ♕e6 mate shows the queen as a bishop on the classical diagonal.) 21 ♗e7+ ♔g6 22 ♘e5+ ♔h5 23 ♕h3 mate. The queen is a queen, moving on the diagonal and mating on the file. 19...♕c4 20 ♖xf7+ ♔g8 21 ♖f8 mate. Note that this position isn’t regarded as important even though it is instructive. First of all it’s of a tactical nature and secondly it’s only given to show the main principle in its simplest form. Another snippet concerning the seventh rank comes from the time when Bent Larsen went so far as to criticise Nimzowitsch for regarding the seventh rank as an element. However, it is indeed an extremely important element in chess and we’ll repeat it again and again whenever the major pieces are involved. Don’t miss the study by Studenetsky from 1939, position 220, and you’ll get the point. One more aspect when dealing with important positions is to observe how they can be transformed to other important positions. For example a multi-piece ending often transposes to an ending with one minor piece each. A typical example is Fischer – Taimanov, position 269. Fischer – Taimanov Candidates match, Vancouver 1971 White to move First we have the multi-piece ending in itself starting at move 25. Fischer – Taimanov Candidates match, Vancouver 1971 White to move Here the exchange of rooks has just taken place. At move 45 White is technically winning the bishop versus knight ending. Both these positions are regarded as important since they contain different ideas as well as different techniques in their respective positions. Another distinction is that between easy and difficult positions. To be able to understand which is the more difficult we need to know the basic one. For example, if you want to learn how to hold the draw in an ending with two rooks each and four pawns versus three on the kingside as in position 273... Tukmakov – Wojtkiewicz Bern 1993 White to move ...you first need to know how to hold the draw in a single rook ending, where you are up against a rook pawn and a bishop pawn, as in Gligoric – Smyslov, position 233. Gligorić – Smyslov Moscow 1947 White to move What is the most effective way to work through the 300 positions? My recommendation is that you look at the diagram and cover up the solution. You then assess the position and try to find the right move, continuation, method, plan, technique or whatever, depending on the specific position you are examining. If you are already familiar with it you go on to the next. If you aren’t familiar with it or need to refresh your knowledge or repeat the position, you can think around 5-30 minutes, depending on your playing strength and overall prior learning of positional ideas. Let’s look at position 86 from a game between Bogoljubow and Alekhine to exemplify in a more concrete form how to get to work on an important position. Bogoljubow – Alekhine World Championship match 1929 White to move The important thing to understand about this position is that White has the initiative in spite of the bishops of opposite colour and the symmetrical formation – a factor which is underlined by the manoeuvre of the queen’s rook from a1-d1-d3-f3 with pressure on f7. This is the most important idea to remember and which essentially led to a win in this World Championship game against Alekhine, who by the way wasn’t particularly fond of defensive play. When you play through the game to the end you will learn other things such as the art of playing over the whole board which was Bogoljubow’s speciality, as well as the psychological importance of having a slight initiative against an attacker. So there are several important ideas to learn from this specific position alone and which you can use in many other positions. When you have acquainted yourself with the key idea(s) you can repeat them after a week to find out how much you actually remember and understand. My advice is that you study five positions each week and then repeat them one week later and work on five more. This is in accordance with the principle “less is more” which means it’s better to do little than nothing, otherwise the risk increases that you’ll achieve nothing if you practice too quickly and too hard. And if you don’t repeat the positions you will forget them. Remember that repetition is at the heart of this pedagogic concept and the need to be pedagogical is your own responsibility in order to maximise the benefits of your training. After all, repetition is the mother of all learning, so don’t forget it if you have serious ambitions to develop your chess understanding as well as your practical playing skills. You will notice that many of the longer variations are supported by smaller diagrams without numbers and the purpose is to make it easier to learn the specific theme(s) visually. Sometimes an interesting pawn ending can show up in another type of ending and then a small diagram presents it graphically. If the pawn ending is regarded as important the diagram has a number attached to it. I wish you good luck with the 300 most important positions and positional ideas and don’t forget that the key to the whole concept is that fewer positions will mean that more positions will be remembered! When you have finished reading the book feel free to contact me at thomasengqvist@protonmail.com. I would really appreciate your comments and feedback for possible future editions. Part 1: 150 most important positions in the Opening and the Middlegame 1-30: Development of the Pieces The key factor in the opening is to understand the value of development. The more open the position, the more important is development. In such situations it’s normal to develop all one’s pieces (including the queen’s rook!) before starting an attack. However, if the opening has a closed character, quick development is usually of less importance. Then manoeuvring will sometimes take precedence. The difficult question is what to do when the game is semi-open or half-closed and everything depends on the concrete nuances in the position. Another important concept is the second wave of development, that is when one or several pieces move twice in the opening after the main development has more or less been completed. This sometimes signifies the transition between the opening and the middlegame, especially if there is a ready target to attack. The first player to understand the importance of swift development in open games was Paul Morphy (1837-1884) and so we start off with him. The following two positions are favourites of mine when I teach my students the value of development, rather than premature attacks or defences, in the early stages of the game. 1 Morphy – Stanley USA 1857 White to move This position arose in the Evans Gambit after: 1 e4 e5 2 ♘f3 ♘c6 3 ♗c4 ♗c5 4 b4 ♗xb4 5 c3 ♗a5 6 d4 exd4 7 0-0 d6 8 cxd4 ♗b6. 9 ♘c3! Morphy was the first player in chess history to regularly play this apparently straightforward developing move of the knight. To this day it is considered the best move. White’s advantage lies in his strong control of the centre and advantage in time. White will have to consolidate these temporary advantages to gain full compensation for the sacrificed pawn. Another great player, Adolf Anderssen, preferred the attacking 9 d5?! but ever since Morphy’s discovery that move was considered superficial. Although it has the advantage that the long darksquared a1-h8 diagonal is opened, the classical a2-g8 diagonal remains blocked while Black’s bishop on b6 becomes stronger. It’s more important to develop pieces instead of making useless short-term attacking or defensive moves. In other words one should avoid unproductive one-move threats and instead concentrate on long-term gains. 2 Meek – Morphy USA 1855 Black to move This position arose after 1 e4 e5 2 ♘f3 ♘c6 3 d4 exd4 4 ♗c4 ♗c5 5 ♘g5?! and now Morphy played the developing move: 5...♘h6!. From his games we can deduce that it’s best to develop the pieces before starting an attack. And this position provides solid proof that Morphy’s principle of rapid development was superior to the style which preferred to attack before carrying out sufficient development. For a less experienced player it might be tempting to play 5...♘e5? which defends and attacks with a single move. However, after 6 ♘xf7 ♘xf7 7 ♗xf7+ ♔xf7 8 ♕h5+ g6 9 ♕xc5 White’s premature attack is suddenly justified due to the fact that Black broke one of the most basic principles in the opening: the principle of development. 6 ♘xf7 ♘xf7 7 ♗xf7+ ♔xf7 8 ♕h5+ g6 9 ♕xc5 d6 and Black is ahead in development and White must fight for equality. 3 P. Morphy – A. Morphy New Orleans 1949 White to move After 1 e4 e5 2 ♘f3 ♘c6 3 ♗c4 ♗c5 4 b4 ♗xb4 5 c3 ♗c5 6 d4 exd4 7 cxd4 ♗b6 8 0-0 ♘a5 9 ♗d3 d5?? (9...d6) 10 exd5 ♕xd5?? (10...♘e7) 11 ♗a3 ♗e6 12 ♘c3 ♕d7 we reach the diagram position. This is a primitive example showing the ground-breaking discovery made by Morphy when he was only twelve years old. When you are ahead in development you should open the central files by a pawn sacrifice. 13 d5! ♗xd5 14 ♘xd5 ♕xd5 15 ♗b5+! Morphy clears the d-file. 15...♕xb5 The central files are opened and it’s time to mate. 16 ♖e1+ ♘e7 17 ♖b1 ♕a6 18 ♖xe7+ ♔f8 19 ♕d5 As mentioned in the introduction, quicker and more spectacular was 19 ♖xf7+! resulting in a king hunt after 19...♔xf7 20 ♕d7+ ♔f6 21 ♗e7+ ♔g6 22 ♘e5+ ♔h5 23 ♕h3 mate. 19...♕c4 20 ♖xf7+ ♔g8 21 ♖f8 mate. White won with the help of the a3-f8 and a2-g8 diagonals and this was typical of the romantic period in chess. 4 Schulten – Morphy New York 1857 Black to move The position arose after the moves 1 e4 e5 2 f4 d5 3 exd5 e4 4 ♘c3 ♘f6 5 d3 ♗b4 6 ♗d2 and now Morphy played 6...e3! opening the central file where the opponent’s king stands. The pawn sacrifice is a pendant to 13 d5 in the previous game. 7 ♗xe3 0-0 8 ♗d2 ♗xc3 9 bxc3 ♖e8+ 10 ♗e2 ♗g4 11 c4 5 Black to move 11...c6! Black wants to open the central files and so it’s logical to attack the spearhead on d5 rather than the base on c4 with 11...b5!?. 12 dxc6 ♘xc6 13 ♔f1 Now that both the central files are opened Black should exploit his more active pieces and look for a decisive combination. 13...♖xe2 14 ♘xe2 ♘d4 15 ♕b1 ♗xe2+ 16 ♔f2 ♘g4+ 17 ♔g1 ♘f3+! 18 gxf3 ♕d4+ 19 ♔g2 ♕f2+ 20 ♔h3 ♕xf3+ 21 ♔h4 ♘h6 22 h3 ♘f5+ 23 ♔g5 ♕h5 mate. My recommendation is that you learn this model game by heart. In particular, moves 6 and 11 are important ideas to remember. When you are ahead in development you must open the central files. When the centre is opened you attack with all the forces you have at your disposal. 6 Wojtaszek – Wang Yue Poikovsky Karpov 2012 Black to move The diagram position arose after 1 d4 d5 2 c4 c6 3 ♘c3 ♘f6 4 e3 ♗f5 5 cxd5 cxd5 6 ♕b3 and now Wang Yue played the revolutionary improvement: 6...♘c6!. The best way to meet threats is by development. Remember the game Meek – Morphy where Morphy played 5...♘h6!. Wang Yue’s choice is the most logical and active move, since it accelerates his development. 6...♗c8 is also playable but passive while 6...♕b6 7 ♘xd5 ♕xb3 loses a pawn due to the intermediate check 8 ♘xf6+. If White plays 7 ♕xb7 a tempo is lost. After 7...♗d7 White loses two more tempi to bring the queen back to a safe place. Black plans ...♖b8 followed by ...♘b4 and ...♗f5. Black’s second idea is to open the position and exploit his lead in development. 8 ♕b3 ♖b8 9 ♕d1 The queen is back home but it has cost time. François-André Danican Philidor (1726-1795) made the interesting statement that a pawn is worth three tempi. If we count all the queen moves, White has made four yet his queen is still on d1. However, we have to deduct one tempo for Black’s bishop move to d7. Now, according to Morphy’s principles, Black should open the game. Remember 6...e3! from the game Schulten – Morphy. After 9...e5 there is no question Black has good compensation since he is ahead in development and has seized the initiative. Apart from that White has an additional problem in finding an active position for his queen’s bishop. 7 Tal – Smyslov Candidates tournament, Yugoslavia 1959 White to move 5 d4! This isn’t a loss of time since the d7-knight hampers Black’s free development. Therefore it’s good for White to open the game. Less precise is the slower 5 g3, since it doesn’t exploit the fact that Black is behind in development on both wings. 5...dxe4 5...exd4 6 exd5 cxd5 7 ♘xd4 leaves Black with a structural weakness on d5, bearing in mind that Black’s pieces are not actively placed. For example, the queen’s knight should be placed on c6 in such a position. 6 ♘xe4 exd4 7 ♕xd4 ♘gf6 8 ♗g5 ♗e7 9 0-0-0 White has a great lead in development and Tal won quickly after: 9...0-0 10 ♘d6 ♕a5 11 ♗c4 b5 12 ♗d2 ♕a6 13 ♘f5 ♗d8 14 ♕h4 bxc4 15 ♕g5 ♘h5 16 ♘h6+ ♔h8 17 ♕xh5 ♕xa2 18 ♗c3 ♘f6 19 ♕xf7! ♕a1+ 20 ♔d2 ♖xf7 21 ♘xf7+ ♔g8 22 ♖xa1 ♔xf7 23 ♘e5+ ♔e6 24 ♘xc6 ♘e4+ 25 ♔e3 ♗b6+ 26 ♗d4 Black resigned. Tal played like Morphy and opened the game as soon as he had realised he was ahead in development. 8 Smyslov – Euwe Candidates tournament, Neuhausen/Zürich 1953 White to move 7 e4! Smyslov makes a sound positional pawn sacrifice to gain the initiative. It may be rather an exaggeration to say that if Petrosian avoided risks and Tal welcomed them, then Smyslov was somewhere in between these players – which speaks for the soundness of the move! 7...dxe4 8 dxe4 ♘xe4 9 ♘d4 ♘xd2 10 ♗xd2 ♗h7 11 ♗c3 More precise was 11 ♖e1 keeping all options open for the darksquared bishop. For example, it could land on a5 if Black places his knight on b6. Nevertheless, White has the advantage in development and both central files are opened. Therefore he has full compensation for the pawn. 9 Alekhine – Tartakower Dresden 1926 White to move White has the more active position and is ahead in development. For these reasons he should open the centre. Alekhine played the principled move: 12 f4!. Another way to open the position, but on the kingside, was 12 h4!? followed by h5. Note that 12 0-0? gives Black the opportunity to lock the position after 12...f4 and slow down the game. This is only to Black’s advantage as he lags behind in development. After 12...exf4 13 0-0 ♘a6 14 ♖xf4 ♘b4 15 ♖h4 and Alekhine won in 55 moves. 10 Timman – Winants 1988 White to move White is ahead in development and so, according to Morphy’s principles, it is important to open the position and seize the moment. Black has helped White in doing this by exchanging on d5. After the strong 11 0-0-0! White’s threat is to invade on d5 with the knight. The routine move 11 cxd5? is a mistake, since after 11...♗f5 Black has caught up in development by having four developed pieces against White’s four. Besides this Black has the initiative, since there are some unpleasant threats White must defend against. 11...♗xc3 12 ♘xc3 ♘xc3 13 ♕xc3! ♕xc3+ 14 bxc3 dxc4 15 ♗xc4 All Black’s active pieces have been exchanged and the d6-pawn is doomed. White won after only nine more moves. 11 Petrosian – Ebralidze Tbilisi 1945 White to move If you are ahead in development and there are no pawn breaks available what are you supposed to do? The reasonable developing move 15 ♖fd1 was played in the game but it wasn’t the strongest continuation. The critical move was 15 c5! in the style of Morphy. White tries to open the position since he is ahead in development. 15...b6 After 15...♕xc5? 16 ♖ac1 ♕e5 17 f4 ♕e6 18 ♔h1 White threatens not only ♖xc7 but also ♕d4 with simultaneous threats against the g7 and a7 pawns. If Black postpones an opening of the game by 15...0-0 White plays 16 ♖ac1 followed by f2-f4 and e2-e4. Sooner or later the c8-bishop must get out and then the position will automatically open up to White’s advantage, as he has more pieces in play. 16 cxb6 axb6 17 f4 followed by placing the rooks on the c- and d-file with strong positional pressure on Black’s position. At higher levels there are many subtleties when in the process of developing the pieces and in which order. Here are two examples, one from the Sicilian Defence and another from the English Opening. 12 Bologan – Frolov Moscow 1991 Black to move After 1 e4 c5 2 ♘f3 e6 3 d4 cxd4 4 ♘xd4 Black can develop the queenside knight with 4...♘c6. This move is Taimanov’s clever refinement of the Paulsen variation as it exerts more pressure on White’s position and limits White’s choices when compared to 4...a6. It saves time and postpones ...a7-a6 to a time when that may become necessary. Kotov named it “the New Paulsen” but eventually the system was deservedly called the Taimanov Variation. Taimanov decided on this move order following his devastating loss against Karpov in Moscow 1972 after playing in Paulsen style with 4...a6 5 ♗d3 ♗c5 6 ♘b3 ♗b6 (Nowadays 6...♗e7 is usual, so as to prepare for a Hedgehog position.) 7 0-0 ♘e7 8 ♕e2 ♘bc6 9 ♗e3 ♘e5 10 c4! ♗xe3 11 ♕xe3 0-0 and now the strongest move was 12 ♗e2! preparing to grab more space with f2-f4. Taimanov’s fourth move is also a clever way to reach the Scheveningen variation without allowing the dangerous Keres Attack after 4...♘f6 5 ♘c3 d6 6 g4. 5 ♘c3 The early attempt to establish a Maroczy Bind pawn formation by 5 c4 is met by 5...♘f6 6 ♘c3 ♗b4 putting more pressure on the white centre and forcing 7 ♘xc6 bxc6 8 ♗d3 (8 ♗d2!?) with a comfortable position for Black, who can choose between 8...d6, 8...e5, 8...0-0 or 8...d5 in the quest for equalisation. The only way to get a reasonable Maroczy position is by playing the time-consuming manoeuvre 5 ♘b5, practically forcing a Hedgehog after 5...d6 6 c4 (6 ♗f4 is pointless according to Karpov. However, one psychological point is that Black is forced to form another pawn structure after the compulsory 6...e5.) 6...♘f6 7 ♘1c3 (After 7 ♘5c3 Black doesn’t need to play ...a6.) 7...a6 8 ♘a3 which is perhaps not to everyone’s taste even though the Hedgehog formation has a good reputation. (8 ♘d4 makes it easier for Black to play 8...d5 immediately or after preparing it with 8...♗e7. 8...♗d7 is also good enough for equalisation.) 13 Black to move The normal move is 8...♗e7, and then 9...0-0, but a more clever move order is 8...b6, preparing a quick ...♗b7 followed by ...♘b8-d7 where the knight is more harmoniously placed than on c6. The knight is better placed on d7 since it’s not in the way on the c-file and can attack either the c4-pawn with ...♘e5 or the e4-pawn with ...♘c5 depending on circumstances. Anand – Illescas Cordoba, Linares 1992, continued 9 ♗e2 ♗b7 10 0-0 ♘b8 11 f3 ♘bd7 12 ♗e3 ♗e7 13 ♕d2 0-0 14 ♖fd1 ♕c7 15 ♖ac1 ♖ac8 16 ♗f1 ♖fe8 17 ♔h1 ♕b8 18 ♘c2 ♘e5 19 b3 ♗a8 20 ♗g1 ♖ed8 with mutual chances. A third option is the Kasparov Gambit 8...d5?!. It was discovered by Kasparov in the spring of 1985 while moving the pieces on a pocket set during a flight. Kasparov played this “super bomb” in the 12th and 16th games in the World Championship match against Karpov during the autumn. The 16th game continued 9 exd5 exd5 10 cxd5 ♘b4 11 ♗e2 ♗c5?! (11...♘bxd5 wins back the gambit pawn but gives White the slightly better game after 12 0-0.) and now Karpov missed the strongest continuation 12 ♗e3! ♗xe3 13 ♕a4+ ♗d7 14 ♕xb4 ♕b6 15 ♕xb6 ♗xb6 16 ♘c4 when Black is facing an unpleasant struggle. 5...d6 6 g4 The Keres Attack obviously has less bite without having a knight on f6 to attack with gain of tempo, but the move is nevertheless purposeful since it’s useful to prepare a pawn storm against the black king’s future residence. 6...a6 7 ♗e3 ♘ge7 14 White to move Black’s position is somewhat lacking in harmony and if it were his turn to move he would exchange the c6-knight and then move the e7-knight to c6. The e7-knight is superfluous, to use Dvoretsky’s terminology, and in reality it’s the e7-knight which wants to be exchanged for the centralised knight on d4. In such a situation it’s sensible for White to put up an obstacle to Black’s plan. 8 ♘b3! Such a move makes Black feel a little uncomfortable. 8...♘g6 Generally speaking, a knight is not well placed on g6 in the Sicilian Defence. 9 ♕e2! The placement of the queen on e2 seems to be lacking in harmony with White’s position but has the advantage that from d1 his queen’s rook will have a good view along the semi-open d-file. Besides, White will avoid threats from one of the knights on the c4square or perhaps f3 in the near future. The queen defends the g4pawn and the light-squared bishop can be developed on g2. The alternative continuation 9 ♕d2 b5 10 0-0-0 ♘ge5 11 g5 ♘a5 12 f4 ♘ec4 gains a tempo against the white queen. After the further 13 ♕f2 ♖b8 14 f5 ♘xe3 15 ♕xe3 ♕b6 Black gains another tempo since White isn’t interested in exchanging queens on Black’s terms. After 16 ♕g3 ♗e7 17 ♘xa5 ♕xa5 18 e5 d5 19 f6 ♗b4 20 ♘xd5 exd5 21 e6 White has compensation for the sacrificed piece as proved by the game Shirov – Polgar, Yerevan 1996. 9...♗e7 10 0-0-0 b5 11 f4 h6 12 ♕f2 ♗d7 15 White to move 13 ♔b1 According to Bologan the white king is placed on the safe squares h1 or b1 in 90% of Sicilian Defence games. Such statistical information proves that there must be a very good reason if White chooses to avoid such a strong prophylactic and consolidating move before launching a real offensive, since once the attack has begun there is no turning back. One might also regard ♔b1 as the completion of queenside castling since the king might turn out to be exposed on the c1-h6 diagonal and in addition the a2-pawn can be vulnerable. 13...♖b8 14 ♗e2 ♘a5 15 ♘xa5 ♕xa5 16 ♗d3 ♗h4 This position is a good example of when there is no need to worry about 16...b4. White simply plays 17 ♘e2 followed by ♘d4 where the knight is very well placed, while the black attack can proceed no further. White is slightly better and won after 41 moves. 16 Marin – Korneev Porto Mannu Open 2008 White to move After 1 c4 e5 it has been discovered by Marin that 2 g3 is probably the most precise move. One advantage with this move order is that it avoids the popular variation 2 ♘c3 ♗b4. For example 3 ♘d5 ♗c5 is a variation favoured by Anand. 2...♘c6 White’s control of the centre has been temporarily weakened so Black can start active play in the centre with 2...c6. This is best answered by 3 d4 exd4 (3...e4 is fully playable and is perhaps best exploited by 4 d5 isolating the far advanced e4-pawn.) 4 ♕xd4. 17 White to move 3 ♘c3! It’s tempting to play 3 ♗g2 to avoid the disturbing ...♗b4 but after 3...f5 (Also interesting is 3...h5 4 ♘f3 e4 5 ♘h4 g5 as mentioned by Nepomniachtchi. The rapidplay game Ivanchuk – Nepomniachtchi, Leuven 2017, continued with the hair-raising 4 d3?! h4 5 g4?! h3! and Black had the initiative.) 4 ♘c3 ♘f6 White has no useful move. 5 e3 (5 d3 ♗b4 gives Black an active game.) is answered by the surprising 5...d5! 6 cxd5 ♘b4 and White can claim no advantage. It’s no coincidence that Black has such a good variation at his disposal, bearing in mind that both of his knights are well developed. The third move played by White is a good illustration of the wellknown principle advocated by Emanuel Lasker which states that knights ought to be developed before the bishops. There are two reasons for this: firstly, knights are stronger than bishops in the early part of the game and, secondly, it’s easier to develop the knights to squares where they belong compared to the bishops which have more squares to choose from. 3...♘f6 3...f5 might be answered by 4 e3, preparing d2-d4. After 3...h5 White can play 4 ♘f3 according to Morozevich. A third option is 3...♗b4 4 ♗g2 ♘ge7 5 ♘d5 ♗c5 6 ♘f3 0-0 7 0-0 d6 8 e3 as in the game Svidler – Nepomniachtchi, Moscow 2018. 4 ♗g2 18 Black to move Sometimes it isn’t easy to decide how to develop the bishop and psychological factors such as personal taste or a particular opponent might influence the choice. 4...♗c5 This is Karpov’s variation in the Four Knights’ variation but another good option is 4...♗b4. Karpov preferred to play this strategically more ambitious continuation when he was up against heavyweights like Korchnoi or Kasparov. Black exerts pressure on the c3-knight and plans to eliminate it at any time, thereby improving his control of the light squares in the centre. This means that the control of the long light-squared diagonal is in danger after a preliminary exchange on c3 followed by ...e5-e4 and/or ...d7-d5. An effective way to avoid this strategical problem is to reply 5 ♘d5. This is a good move favoured by intuitive players like Karpov, Tal and Gheorghiu. 5 ♘f3 d6 6 d3 0-0 7 0-0 Black’s c5-bishop cooperates well with the pawn chain c7-d6-e5(f4). This is the same kind of cooperation White’s g2-bishop enjoys alongside the pawn chain e2-d3-c4-(b5) and sometimes even a6! The goal for both players is to improve the activity of their bishops. 7...a6 8 a3 ♘d4 9 ♘e1 h6 10 b4 ♗a7 11 e3 White follows the main plan which is to confine Black’s darksquared bishop to passivity. 11...♗g4 12 f3 ♗d7 13 a4 c6 14 a5 ♖e8 15 ♔h1 ♘f5 16 ♘c2 d5 17 e4 ♘d4 18 c5 ♘b3 and a draw was agreed. Note that it’s harder for Black to activate the c- and f-pawns because of the knights that stand in their way, White’s attack on the queenside is normally faster than Black’s on the other flank. A good introduction to learn how to play this variation is to study Karpov’s games with Black where he developed the bishop to both c5 and b4. We will continue the discussion on giving consideration to the development of the bishop(s) with the help of a few instructive examples. We will also touch on the subject of when it’s advisable to prevent the development of the opponent’s bishop. 19 Capablanca – Janowski San Sebastian 1911 White to move In My Chess Career Capablanca writes that his move 13 ♗e2 was a mistake. He realised that 13 g3 was the correct move but was afraid to be criticised for constructing such a weakened pawn formation on the kingside. Because of this reason alone Capablanca opted to go against his better judgement. He points out that it was his first major tournament and he didn’t want to be scolded for making pointless pawn moves which is something weak players are prone to do. 13...♗e6 14 ♗f3 It’s better to maintain the bishop on the e2 square. To play an extended fianchetto (♗f3 instead of ♗g2) has the drawback that it’s easier to attack it with a knight on e5. However, it has the advantage of not having to make a weakening pawn move like g2-g3. 14...♖a7! In comparison to Capablanca, Janowski isn’t afraid of playing “strange” moves! 15 0-0 ♖c7 16 ♕b3 Capablanca considered that 16 ♖c1 was better. 16...♘bd7 17 ♖fd1 17 ♘xd5 or 17 ♘c6 don’t work according to Capablanca. 17...♘e5 18 ♗e2 White has lost two moves with the bishop. 18...♕e7 19 ♖ac1 ♖fc8 20 ♘a4 ♖xc1 21 ♖xc1 ♖xc1+ 22 ♗xc1 ♘e4 23 ♗b2? Better was 23 ♘xe6 with an equal game. For the first time in his life Capablanca felt he had been outplayed by his opponent up to move 23. 23...♘c4 Capablanca and Janowski missed the winning move 23...♕f6!. After 24 ♗f3 ♘d2 25 ♕d1 ♘exf3+ 26 ♘xf3 ♘xf3+ 27 gxf3 (27 ♕xf3? ♕h4 wins a piece.) 27...♕g6+ 28 ♔f1 ♗xh2 Black wins the h-pawn and creates a deadly passed pawn on the h-file. After his 23rd move Capablanca managed to turn the tables and after 66 moves he emerged victorious. It’s noticeable that the pawn would have been better placed on g3 to prevent threats against h2. Moreover if the bishop had been developed to g2 earlier in the game it would have exerted pressure on the d5-pawn. What’s interesting about the whole story is that if you check the opinion of Komodo 11 it actually prefers 11 ♗e2 and thinks it’s the best move and offers a clear advantage (0.94). 11 g3 is valued the third best move (0.81), so maybe Capablanca was too harsh on himself. 20 Keres – Fine Ostende 1937 White to move According to Keres White has two fundamental plans: The first one is to exploit the pawn majority in the centre and play for d4-d5 establishing a passed pawn. The second is about concentrating all his pieces on the black kingside and in that way utilising the centre and his space advantage. In the game Keres played the flexible 11 ♗c4 which is the standard move in this position. Compared with 11 ♗e2 which keeps the d4-d5 push in mind, or 11 ♗d3 planning a kingside attack, the advantage of Keres’s move is that both options are still open. All the same it seems that Keres has forgotten one typical plan which has been used by Petrosian. That is to play for a minority attack on the queenside with a4-a5, especially when Black has played ...b6. 11...♘d7 Black intends the manoeuvre of the knight to f6 or f8, defending the kingside. 11...♘c6 plans to generate play on the queenside. The continuation 12 0-0 b6 13 ♖fd1 ♗b7 14 ♕f4 ♕f6! 15 ♕e3 ♖fd8 16 e5 ♕h6! gave Black good play in the game Reshevsky – Fine, Hastings 1937. Note how persistent Black is in making a queen exchange. If White plays 17 ♕xh6 gxh6 then strictly speaking Black’s splintered kingside pawns are not to be regarded as weaknesses since it’s not possible for White to attack them. The most important factors are that Black has good play in the centre as well as on the queenside. A plausible continuation then would be 18 ♖ac1 ♖ac8 19 ♗b3 ♘b4 with advantage to Black. 12 0-0 b6 13 ♖ad1 13 ♖ac1 is useless in the long run, since it will be easier for Black to endeavour to exchange rooks along the c-file. 13 a4!? is the plan preferred by Petrosian, and also by Yusupov and Beliavsky, to attack the b6-pawn with a4-a5 and create a weakness either on b6 or a7 if Black plays ...bxa5. 13 d5?! is a positional mistake due to 13...♘c5 and the black knight has an excellent position. 13...♗b7 14 ♖fe1 The advantage of placing the rooks behind the classical centre (pawns on e4 and d4) is that White can play for both d4-d5 and e4e5 depending on the circumstances. 14...♖c8 15 ♗b3 Black must now make the difficult decision whether to defend the kingside with 15...♘f6, where the knight is more active but also more vulnerable, or 15...♘f8 where it is safer but more passively placed. 21 Bronstein – Averbakh Moscow 1962 Black to move Capablanca’s recommendation is to wait with the development of the c8-bishop and instead improve the placement of the king, rook, knight and dark-squared bishop. 10...0-0 10...♗e6?! was played in the game Capablanca – Eliskases, Moscow 1936, but Black had problems after 11 ♗xe6! fxe6 12 ♕b3 ♕c8 13 d4 (13 ♘xe5? ♗xe3! Capablanca) 13...exd4 14 ♘xd4 ♗xd4 (14...e5 15 ♘e6) 15 cxd4 and White remained with the freer and more active game. 11 0-0 If 11 d4 then 11...exd4 12 cxd4 ♗b4+. 11...♘g6 12 d4 ♗b6 It’s important to support the e5-pawn so Black isn’t forced to give up the centre with ...exd4 as in the Capablanca game. 13 a4 c6 14 dxe5 dxe5 15 ♕xd8 ♖xd8 16 ♖ad1 ♖e8 17 ♖fe1 ♔f8 18 ♘f5 ♗c7 19 ♗b3 ♘f4 20 ♘e3 a5 21 ♗c2 ♗e6 ...and Black is slightly better although the game ended in a draw after 41 moves. Black displayed good timing with the development of his queenside bishop. Other players who were very skilful with regard to how and when to develop this bishop, were Lasker and Réti. It’s a matter of developing the pieces correctly and in the right order, while at the same time never losing focus on the centre – the most important area of the board. 22 Nepomniachtchi – Nakamura Wijk aan Zee 2011 Black to move Sometimes it’s more important to disturb White’s harmony than to concentrate on development. Here Nakamura came up with the fantastic idea 7...♗g4!? where Black sacrifices a tempo to provoke the pawn move to f3. Svidler – Nakamura, Amsterdam 2010, had continued with the inferior 7...dxc4 8 ♘g3 ♗g6 9 ♗g5 ♕b6 10 ♕d2 ♕b4 11 a3 ♕b3 12 ♘ge4 ♘d5 13 ♖h3! and so Nakamura didn’t want to repeat the variation. One idea behind Nakamura’s new move is simply to block the third rank for the king’s rook. 8 f3 Another drawback of 8 f3 is that the light-squared bishop can no longer exert pressure on the h5-pawn by ♗e2. 8...♗f5 9 ♘g3 ♗g6 10 ♗g5 ♕b6 11 ♕d2 ♘d7 12 a3?! f6 13 ♗e3 ♕b3! Compared to the game against Svidler the black queen is safe on b3. 14 cxd5 ♘xd5 15 ♘xd5 ♕xd5! Black has equalised and managed to win the game after 44 moves. An impressive novelty! 23 Mnatsakanian – Petrosian Moscow 1964 Black to move At times it’s more important to prevent the opponent from developing his bishop rather than developing your own. Here Petrosian played the clever 11...♘d5! to exploit the fact that White has been left with a rather passive dark-squared bishop because of making too many pawn moves. Now the active developing move ♗g5 is effectively prevented. Petrosian’s move is illuminating, since it shows that one must always be aware of the opponent’s development as well. After 12 0-0 Black continued his development with 12...♗d6! and obtained a good game. 24 Winter – Alekhine Nottingham 1936 Black to move Alekhine prevented his opponent from developing the c1-bishop with the move 6...♕h4!. It’s important to stop White’s bishop coming to f4 with a simplified game. There followed 7 ♘d2 ♗g4! 8 ♕c2 (8 ♕b3 ♘ge7) 8...0-0-0 with a good game which resulted in a victory for Black after 39 moves. When one has the bishop pair it sometimes makes sense to open the position rather than develop the pieces. Here are two examples on that theme. 25 Verlinsky – Alekhine All-Russian Championship 1909 Black to move Black has the bishop pair so it’s in his interest to open the position. This explains: 9...c4! 10 bxc4. White has to accept the pawn sacrifice because otherwise Black gets rid of his doubled pawn and still has the initiative thanks to his bishop pair. 10...♗a4? 10...0-0-0 was correct. If Black is unable to win back the pawn, he would still have compensation in view of White’s damaged queenside pawns. 11 c3? Correct was 11 ♘bc3 ♗b4 (11...♗xc2?? 12 ♔d2) 12 ♖b1 ♗a5 13 ♖b2 with a slight advantage to White. Black’s bishop is misplaced on a4. 11...0-0-0 12 ♘d2 ♗c2! 13 f3 ♗c5 and Black won after 28 moves. 26 Karpov – Timman Netherlands 1981 White to move 12 f4! 12 d6? is obviously wrong due to ...♘c6 and ...♘d4. Most players would probably have played 12 ♕c2 to protect the b2-pawn before making the pawn break f3-f4. However, the problem is that White’s bishop would then have problems after 12...f5 13 f4 e4. 12...exf4 12...e4 is answered by 13 f5!. 13 ♗xf4 ♗xb2 14 ♖b1 ♗f6 Or 14...♗d4 15 ♗d6. 15 ♕a4! White has good compensation for the pawn in his active play. In some openings it’s tricky to know where the rooks will be most effectively placed and that may depend to a certain extent on the style of the players. 27 Larsen – Petrosian Piatigorsky Cup, Los Angeles 1966 White to move This position arose after 1 e4 c5 2 ♘f3 ♘c6 3 d4 cxd4 4 ♘xd4 g6 (Accelerated Dragon) 5 c4 ♗g7 6 ♗e3 ♘f6 7 ♘c3 ♘g4 8 ♕xg4 ♘xd4 9 ♕d1 ♘e6 10 ♕d2 d6 11 ♗e2 ♗d7 12 0-0 0-0. An aggressively inclined player would prefer 13 ♖ad1. The main idea of the set-up is to develop the king’s rook by means of the pawn push f2-f4 and, if possible, f4-f5. This aggressive placement of the rooks focuses on the centre and the kingside and suits attacking players much better than the development of the rooks to c1 and d1. Keres – Petrosian, Candidates Tournament 1959, continued more positionally with 13 ♖ac1 ♗c6 14 ♖fd1 ♘c5 (14...♗xc3 15 ♕xc3 ♗xe4 16 ♗h6 or 16 c5 with good play for the pawn.) 15 f3 a5 and Black could be satisfied with the opening. Keres, like Larsen, was well-known for his fondness for attacking play but this time he chose quieter play despite the fact that a sharper continuation was available. This is perhaps one example of why Keres and Spassky were called universal players because they felt at home in both types of games. 13...♗c6 14 ♘d5 ♖e8? 14...♘c5, with the idea 15 ♕c2 a5 16 ♗xc5 dxc5 17 ♘f6+ ♗xf6 18 ♖xd8 ♖fxd8, was a fully playable continuation but for some reason Petrosian avoided it. 15 f4 and Larsen won brilliantly after 30 moves, finishing off with a queen sacrifice – which proves just how dangerous this position is for Black! 28 Karpov – Korchnoi London 1984 White to move Moves such as 13 ♕b3, 13 ♖b1 or 13 ♖c1 look like normal means of development but Karpov thought more deeply than that. He played 13 ♖e1! and realised that the key to maintaining the balance is flexibility. The move Karpov made is useful whatever happens in the future. This is positional play at a high level and the move that maintains optimal tension in the position. Karpov’s move has the intention to meet a later ...f5-f4 with exf4 or possibly e2-e4. If Black plays ...♘f6-e4 White can move the c3knight and plan a future f2-f3. Then the rook is correctly placed on e1 since the e3-pawn is defended and e2-e4 can be prepared. It’s also possible for Black to play ...c6 to meet b5 with ...c5. After dxc5 Black wants to play ...d5-d4 and then the rook will also be well placed on e1. It also makes sense to keep the queen on d1 (and avoid ♕b3) in case Black plans a kingside attack. The game continued: 13...g5 14 ♖c1 ♔h8 15 ♗d3 c6 16 b5 g4 17 ♘d2 c5 18 dxc5 ♘xc5 19 ♘b3 ♘xb3 20 axb3 ♖c8 (20...d4 21 exd4) 21 ♘e2 ♖xc1 22 ♕xc1 ♕b6 23 ♘f4 ...and Karpov had obtained a slight advantage which he converted to a win. It’s a good idea to compare Karpov’s 13th move with that which Kasparov played against Karpov in the deciding final game of the World Championship in 1985. You will find this game in the section on manoeuvres. Normally we move each piece only once in the opening. If we move any of them a second time after completing development we can talk of a second wave of development. 29 Rotlewi – Rubinstein Lodz 1907 Black to move The above position is completely symmetrical except for the fact that Black has a rook on d8 and is to move. This is actually one of the first games showing how to benefit from a slight lead in development (two half moves) in a symmetrical position. Rubinstein played 15...♘e5!. The best continuation is to start a second wave of development and take the initiative as soon as possible. The routine 15...♖ac8, completing his development, might have worked as well but then we would hardly have experienced the wonderful combination at the end of the game. 16 ♘xe5 ♗xe5 All Black’s minor pieces are a little better placed than those of his opponent because White is weak on h2 and g2 and the f6-knight can more easily go for an attack compared with White’s knight. 17 f4?! The developing move 17 ♖fd1 was preferable. 17...♗c7 18 e4?! ♖ac8 19 e5? Better is 19 ♖ad1 ♗b6+ 20 ♔h1 h5. 19...♗b6+ 20 ♔h1 ♘g4! 21 ♗e4 If 21 ♕xg4 ♖xd3. 21...♕h4 21...♘xh2!? was also good but who would criticise Rubinstein for not playing it after seeing the immortal continuation? 22 g3 22 h3 ♖xc3! 23 ♗xc3 ♗xe4 24 ♕xg4 (24 ♕xe4 ♕g3 25 hxg4 ♕h4 mate.) 24...♕xg4 25 hxg4 ♖d3. 22...♖xc3!! 23 gxh4 23 ♗xc3 ♗xe4+ 24 ♕xe4 ♕xh2 mate. Or 23 ♗xb7 ♖xg3. 23...♖d2!! 24 ♕xd2 ♗xe4+ 25 ♕g2 ♖h3!! The queenside rook decides the game in the most beautiful fashion. White resigned. Apart from the fact that this is an enormously beautiful and rare combination, demonstrating the dynamic potential of the pieces, it is an important model game showing how to play for a win in a symmetrical position. 30 Matanović – Petrosian Kiev 1959 Black to move 10...♗g4! A clever prophylactic move played at the last moment to eliminate any idea of a second wave of development with 11 ♘e5 which would give White some chances of claiming the initiative, for example after 10...0-0?! 11 ♘e5. Another more direct idea was to focus on the light squares and play 10...b5 11 ♗d3 ♗xd3 12 ♕xd3 ♕d5 13 ♔b1 0-0 with mutual chances. 11 h3 ♗xf3 12 ♕xf3 It looks like White has won the bishop pair for nothing but after 12...♘d5 the dark-squared bishops will be exchanged by force. 13 ♗xe7 White cannot avoid the exchange if Black persists, since 13 ♗d2 can be answered by 13...♗g5. 13...♕xe7 14 ♖he1 0-0 15 ♔b1 ♖ad8 ...and the position is about equal 31-64: Exchange of Material Most exchanges of material affect the evaluation of a position one way or another but some exchanges definitely do, such as bishop for knight, bad bishop for good bishop, four-rook-swap and, above all, the exchange of queens. The last mentioned exchange often signifies a transition from middlegame to endgame. We’ll look at all the individual pieces but first here is an example showing how and when to exchange pawns. 31 Capablanca – Alexander Nottingham 1936 White to move 27 g4! This is how to get rid of the tactical g3-weakness. Notice how White exchanges his worst pawn against Black’s good one. 27...♕g6 28 gxf5 ♖xf5 The ugly g3-pawn has been exchanged for the pretty f5-pawn and Black is now saddled with a tactical weakness on g5. After the inferior 28...♕xh6? White gets rid of his e3-pawn by 29 e4!. 29 ♖xf5 ♕xf5+ 30 ♔a1 White is clearly better since Black has a weak pawn on d5 and the kingside has also been weakened. The g5-pawn would have been better placed on g6. 30...♖f8 32 White to move 31 ♕h1! This kind of hypermodern queen manoeuvre is an inspiration from the great Réti! Here the queen is very well placed, since it not only exerts pressure on d5 but also protects h6 and prepares the pawn push to e4. Other ideas are to follow up with ♕h2 and ♖f1. Such a move is grandmasterly due to the multiple ideas involved. 31...♕d3? 31...♗d6 was preferable although White is clearly better after 32 e4!, when White gets rid of his annoying pawn by exchanging it for Black’s bad d5-pawn. 32 ♘b1 ♗d6 33 ♗f3? White leaves room for the rook on g1 to put pressure on g5 but 33 ♗b4! ♗xb4 34 ♖xc7 ♗a8 35 ♖xa7 ♕xe3 36 a3 ♗e7 36 ♗xd5 was better, winning on the spot. 33...♕f5 34 ♖f1 ♕g6 35 ♖g1! White plans ♘c3 and e4 with pressure on Black’s weak pawns, especially the one on g5. 35...♗e7?! A more active defence was 35...♕f5. 36 ♘c3 36...♕g8? Alekhine regarded this as a bad move and it’s certainly passive. Black could try Prins’s suggestion 36...♕e6 (Or perhaps the unhuman computer choice 36...a5!?.) 37 e4 dxe4 38 ♗xe4 ♗xe4 39 ♕xe4 ♕xe4 40 ♘xe4 and now 40...♖g8! (Instead of 40...g4 41 ♔b1 ♖g8 42 ♗f4 winning according to Prins.) with some drawing prospects. For example, 41 ♘xg5 (41 ♗xg5? ♘e6 42 ♗f6+ ♗xf6 43 ♖xg8+ ♔xg8 44 ♘xf6 ♔f7 45 ♘xh7 ♔g6 46 d5 ♘d8 47 ♘f8+ ♔xh6 is a draw.) 41...♗xg5 42 ♗xg5 ♖g6 43 ♔b1 ♘d5 and ...♔g8 and ♔f7 with good control of the white squares. 37 e4! This move is proof that White’s strategy has been successful. White has not only created a passed pawn but also liberated his dark-squared bishop. 37...♕f7 38 ♗g2 ♕f2? Better was 38...dxe4 39 ♘xe4 with a superior position for White. 39 ♗xg5! The game is over. 39...♗xg5 39...♗b4 loses after 40 ♕h2!. 40 ♖f1 ♕xf1+ 41 ♗xf1 dxe4 42 ♗g2 ♘e6 43 ♗xe4 ♗a6 44 ♗b1! Black resigned. Let’s move on and see how to deal with tensions between knights. 33 Petrosian – Ebralidze Tbilisi 1945 White to move The young Petrosian played the routine move... 9 ♗g2?! ...but correct was 9 ♘c3!. White has a space advantage in the centre and should avoid unnecessary minor piece exchanges. In the game followed: 9...♘xe4 10 ♗xe4 and then another opportunity to exchange arose for Black after 10...♗b4+ 11 ♗d2 ♕e7. Black had a comfortable game but Petrosian managed to win anyway. Remember that a space advantage often decreases in importance for every piece exchange that’s made. 34 Botvinnik – Zagoriansky Sverdlovsk 1943 White to move Even a seemingly harmless exchange of a pair of knights can lead to a substantial advantage. Here Botvinnik played 19 ♘e5! forcing the exchange of the last pair of knights. It results in permanent weaknesses on the dark squares in the centre, particularly d4. The d4-square will be an important springboard to enter Black’s position. The move played also vacates the f3 square for the bishop to put the d5-pawn under pressure. 19...♘xe5 20 ♕xe5 Black’s remaining pieces, with the exception of the queen, are located in passive positions so as to defend the isolated pawn. Black also remains with a positionally bad bishop. 20...♕c5 21 ♗f3 b6 22 ♕b2 ♖c8 Black holds on to the c-file and that is some consolation for being in possession of an isolated pawn with all the best minor pieces exchanged. 23 ♕e5 White forces Black’s rook back to a defensive position. 23...♖cd8 After 23...♕c3 White wants to keep the queens on the board with 24 ♖d4. Inaccurate is 24 ♕xc3 ♖xc3 25 ♗xd5? ♖xd5 26 ♖xd5 ♗xd5 27 ♖xd5 ♖c1+ 28 ♔h2 ♖c2 with a draw. 24 ♖d4 a5 The continuation can be seen in the section Manoeuvring with Several Pieces, position 107. 35 Fischer – Petrosian Buenos Aires 1971 White to move Here Fischer played the surprising move 22 ♘xd7+!. Fischer, like Capablanca, was a true master when it came to the conversion of advantages. By exchanging Black’s “bad bishop,” which was about to be liberated after ...♗b5, White can in the near future use the queenside rook to penetrate the black position via the c6- or the c7-square. Note that the exchange also represents a transition phase between the middlegame and the endgame, according to Speelman’s definition that was mentioned in the introduction. The routine move 22 a4 gives Black better defensive resources after 22...♗c6 followed by ...♘d7. 22...♖xd7 23 ♖c1 White’s advantageous trade also opened the c-file and in combination with Black’s pawn weaknesses on a6 and d5 these factors will prove to be decisive. 23...♖d6 Black stops White from playing the rook to c6 but now the rook comes to the seventh rank instead. 24 ♖c7 ♘d7 Black must prevent the other rook from invading on e7 where it would be a mighty force in conjunction with its colleague on c7. To place two rooks on the seventh rank is the ultimate dream for two healthy rooks. 25 ♖e2 g6 It’s hard to find constructive moves for Black. 26 ♔f2 Fischer makes use of the no-hurryprinciple and exploits the time to progressively improve his position. 26...h5 27 f4 All black pawns are on light-coloured squares and that is naturally good for the white bishop. 27...h4 28 ♔f3 f5 29 ♔e3 d4+ Black wants to prevent ♔d4 but weakens the classical a2-g8 diagonal. 30 ♔d2 ♘b6 Black stops ♗c4 but allows… 31 ♖ee7 ♘d5 32 ♖f7+ ♔e8 33 ♖b7 ♘xb4 34 ♗c4 and Petrosian resigned. White is attacking the knight while also threatening the decisive ♖h7. 36 Carlsen – Anand Linares 2009 White to move Perhaps Carlsen, who is well acquainted with the classics, was inspired by Fischer’s famous exchange when he went for an ending with three pieces each after the surprising 18 ♘c7! ♖ag8 19 ♘xe6!. It might seem illogical to exchange in this way but even a bad bishop can have important defensive functions such as protecting pawns. The Polish grandmaster Krasenkow, when annotating the game for Chessbase, wrote: “A good idea. White exploits the fact that Black cannot recapture with his f7 pawn to exchange the theoretically “bad” e6 bishop, which in fact is the main defender of the black pawns.” 19...♔xe6 20 ♖xg8 ♘xg8 21 ♔e2 ♘e7 22 ♔f3 One can clearly see that Black has three positional weaknesses on f5, d5 and f7, which result from the fact that the important defensive e6-bishop is gone. Besides, in general the combination of rook and bishop versus rook and knight is a little stronger when there are pawns on both sides of the board – and here many of Black’s pawns are fixed on white squares! At the right moment the king can advance to f4 and commence an attack on the black pawns. Carlsen won the endgame after 77 moves. 37 Kasparov – Short Brussels 1986 White to move 16 ♖a1! This is the correct move since the white knight wants to go to e5 and exploit the strength of the isolani. 16 ♘xg5 ♕xg5 would only be in Black’s interests since he wants to exchange pieces in view of his better pawn structure. 38 Staunton – Williams London 1851 White to move One of the main differences between the long-range bishop and the short-range knight is that often the bishop has more possibilities of exchanging a knight than vice versa. This is not only due to its long-range capability but also because of its ability to pin. 14 ♗xf6!? Creative technique! In most cases it’s regarded as bad to exchange a fianchettoed bishop for a knight, but here it is of definite interest since it weakens Black’s position in the centre. Many players would probably prefer the standard move 14 ♗a1 which is not bad but rather a sign of pedestrian technique. Incidentally, the computer suggests the highly creative moves 14 d4 or 14 ♖b1. 14...♕xf6 15 cxd5 exd5 16 d4 The idea of White’s exchanges is to apply strong pressure against Black’s pawn centre. 16...c4? 16...cxd4 was better but White stands rather better after 17 ♘xd4. 17 ♘e5 and Staunton won quickly after only 32 moves. 39 Georgescu – Stein Bucharest 1961 Black to move 16...♗xc3!? This is a surprising decision by Stein, who showed that he was far from being a dogmatic player wanting to keep the bishop pair at all costs. Another player who doesn’t mind parting with the bishop pair is the reigning World Champion Magnus Carlsen. What is more astonishing about Stein’s decision is the fact that he has worked hard to get a strong bishop on g7 only to exchange it for a knight which looks pretty harmless on c3. 16...♗e8 was playable but passive and probably not suited to Stein’s active style. 16...♘a5 17 ♕xd6 ♘xb3+ 18 axb3 ♕xd6 19 ♖xd6 hxg5 20 fxg5 (20 hxg5 ♖xh1+ 21 ♗xh1 ♖h8 22 ♖d1 ♗c6) 20...♗e5 21 ♖d3 a5 offers good compensation for the pawn since the bishop is strongly placed on e5. Perhaps Stein was averse to an exchange on b3 on principle, since this was something he could have avoided even if it were a bishop that was placed on b3. 17 ♕xc3 e5 This is the main idea of the exchange since the danger of a possible knight move to d5 has now been eliminated. 18 fxe5 18...♘xe5! Here the knight is strong since it also defends the weakness on f7. 18...dxe5 is playable but leads to a more passive position after 19 ♖df1 ♗e8 or 19...♗e6 which is answered by 20 ♗h3. 19 ♖xd6? 19 ♕d2 was better. 19...♕xc3 20 bxc3 hxg5 21 hxg5 ♖xh1+ 22 ♗xh1 ♗c6 23 ♖d1 ♖h8 24 ♘d4 ♖h3 Black threatens to capture either on c3 or g5 after ...♖g3. Black won after the further: 25 ♖g1 ♖xc3 26 ♗g2 ♔c7 27 ♘xc6 ♔xc6! 28 ♗f1 ♔c5 29 ♗d3 ♘xd3+ 30 ♔d2 ♔d4 31 ♖g3 ♘c5! and White resigned. 40 Botvinnik – Tal World Championship match, Moscow 1961 White to move Here the exchange 23 ♗xf6! gxf6 and the follow up 24 b4 can be regarded as an example of routine technique in its most positive sense and in the spirit of Capablanca. White´s bishop exchange weakens Black’s pawn structure and the further goal is to strive for a four-rook-swap to remove all possible counterplay with the heavy pieces. Botvinnik’s annotation here reads: “Black has been left with weak pawns, and the ending (with or without the rooks) will be difficult for him.” 24...♗f5?! Botvinnik recommended 24...♗d5 when the bishop dominates the knight. It’s surprising that Tal didn’t display more persistence and choose to struggle against the knight but presumably he had some tactical ideas in mind. 25 ♘b3 ♗d3+ 26 ♔e1 26...b6!? Black wants to dominate the knight but the problem is that the c6-pawn becomes a weakness. It’s probably better to exchange one pair of rooks by playing 26...♗f5 although Black’s position remains uncomfortable. 27 ♖ac1 Maybe Botvinnik refrained from 27 ♘d4 due to the imaginative 27...♗f1!?. After 28 ♔xf1 c5 29 bxc5 bxc5 30 ♖d3 cxd4 31 ♖ad1 White wins a pawn, but in a rook ending it’s sometimes harder to convert this. Another possible scenario is 27...♗e4 28 f3 c5 29 bxc5 bxc5 30 ♘b3 ♗c2 31 ♖xd7 ♖xd7 32 ♘xc5 ♖d5 33 ♘a6! where it’s perhaps not so easy to exploit the extra pawn. 27...♗e4 28 f3 ♖xd1+ 29 ♖xd1 ♖xd1+ 30 ♔xd1 ♗d5? It was decidedly better to play 30...♗f5 followed by ...♗d7. This knight versus bishop ending will be analysed in the endgame section, position 199. 41 Smyslov – Letelier Venice 1950 White to move 17 ♗xf5! It’s the correct decision to exchange the light-squared bishop for Black’s knight, thereby blocking the f-file which Black was ready to open. 17...exf5 18 exf6 ♖xf6 19 ♖ab1 White has the advantage owing to the weak dark squares on e5 and c5. 19...h6 20 ♖b5! ♗e6 21 ♖eb1 ♖ff7 22 ♘e1 f4 23 f3 g5 24 ♘d3 White’s pieces are located on perfect squares. 24...♔h7 25 ♖e1 ♖f6 26 ♖c5 Smyslov prepares ♘b4. 26...♖c8 26...b6 27 axb6 axb6 28 ♖b5 ♖b7 29 ♘c5. 27 ♘b4 27...♘xb4 27...♘e7 28 ♘xd5! ♘xd5 29 ♖xe6 ♖xc5 30 ♖xf6 ♖xc3 (30...♘xf6 31 dxc5 ♘d7 32 ♔d3! ♘xc5+ 33 ♔c4 and White’s queenside pawns are lost.) 31 ♖d6 A nice example of tactics and technique in the endgame! 28 ♖xe6! ♖xe6 29 ♖xc8 ♘c6 30 a6! 30 ♖c7+ is met by 30… ♖e7. 30...bxa6 31 ♖c7+ ♔g6 32 ♖d7 ♘e7 33 ♗b4 ♘f5 34 ♖xd5 Now it’s just a matter of technique with two passed pawns in the central part of the board. 34...♘e3 35 ♖d8 ♘xg2 36 d5 ♖b6 37 ♗c5 ♖b7 38 ♖c8! ♘h4 39 ♔e2 ♘f5 40 ♖c6+ ♔h5 40...♔f7 42 ♔d3. 41 d6 ♖d7 42 ♖c7 Black resigned. 42 Flohr – Bondarevsky Leningrad/Moscow 1939 White to move One of the overall ideas when playing against the Stonewall is to exchange all the minor pieces except Black’s light-squared bishop and keep one knight which can be placed on the weakened e5square. In this position we can see that White has already exchanged the dark-squared bishop and that exchange has reduced Black’s attacking chances on the kingside. Flohr wrote that in his practical experience an e5-knight will prove stronger than a lightsquared bishop, which to a large extent will be restricted by its own pawns. 21 ♗xd5! exd5 Flohr planned to answer 21...♖xd5 with 22 ♖c5. 22 ♖c3 f4 Forced, otherwise White would have played f4 himself and then Black’s position would have had no future at all. 23 gxf4 23 ♕d2!? was an option. 23...♕h4 24 ♕g3 ♕xf4 25 ♕xf4 ♖xf4 26 e3 ♖f6 27 ♖b3 Flohr wants to lure the rook to b8 so Black’s bishop cannot come out via the d7-h3 diagonal. 27...♖b8 27...b6 weakens c6 and gives White the possibility of doubling on the c-file and tieing down Black’s pieces to the defence of the c6pawn. 28 f3 ♔f8 29 ♔f2 ♖h6 30 ♔g3 Flohr makes good use of the e5-knight to activate the king since Black has no check on g6. Black was under severe pressure and eventually had to throw in the towel after 77 moves. The idea to remember when playing against the Stonewall is to exchange the right pieces. The next example will also show this. 43 Schlechter – John Barmen 1905 White to move If you haven’t seen the following idea before, it’s easy to overlook or underestimate. The objective is to dismantle the Stonewall structure with 14 ♗xe4! fxe4 followed by the pawn break 15 f3 opening the e-file and at the same time getting rid of the doubled f-pawn. 15...exf3 16 ♖ce1 This method of dealing with Black’s centralised knight is well worth remembering since it’s more effective than playing the traditional f2-f3 to drive away the e4-knight. Remember that White’s bishop isn’t necessarily stronger than Black’s as his pawns also help to restrict the white bishop. For this reason alone it makes sense to keep knights on the board as they can later benefit from the weak dark squares. 16...♕c7 17 ♕a3 A wonderful diagonal for the queen. 17...♔g8 Of course not 17...♘d7 when after 18 ♕e7! White’s queen dominates the black position. 18 ♖xf3 ♘a6 19 b3 The queen prepares to manoeuvre to other parts of the board via b2. 19...♕d8 20 c5 ♘c7 21 ♕b2 ♗d7 22 ♕c2 ♕e7 23 ♖ef1 ♖ae8 24 g4 ♗c8 25 ♖h3 g6 White has now forced further weaknesses on the dark f6 and h6 squares. It’s instructive to see how Schlechter exploits them. 26 b4! Schlechter preferred to play over the whole board – as did Bogoljubow and Alekhine a little later. White is superior in the centre and on both wings and can choose whether to break through on the kingside or on the other flank. Therefore it’s good to create the conditions for a breakthrough with b4-b5 while preparing the same situation on the kingside. It’s impossible for Black to defend against two attacks on two different fronts due to his cramped game. In this way White takes advantage of his space advantage as well as Black’s bad bishop. 26...♕f6 27 ♖hf3 ♖e7 28 a4 a6 29 ♘d1! This knight is on its way to g4, after a preliminary g4-g5, in order to exploit one of the holes on g6 or f6. 29...♖g7 30 ♘e3 ♕e7 31 g5 ♗d7 32 ♘3g4 ♗e8 33 ♘h6+ ♔h8 34 ♕e2 White plans to place the e5-knight on f6 and the queen on e5. 34...♕d8 35 ♘eg4 ♗d7 36 ♕e5 ♘e8 37 ♖h3 ♕c7 37...♕e7 is answered by 38 ♕b8!. Note how White exploits the very strong potential of a powerfully centralised queen. 38 ♘f6! 38...♕xe5 Strong pieces have a tendency to be exchanged as Nimzowitsch once said and this is also the case here. Playing a waiting game with 38...♕d8 doesn’t work on account of 39 ♘xh7!. 39 fxe5 ♖e7 40 ♖hf3 ♘xf6 41 ♖xf6 ♖xf6 42 exf6 ♖e8 43 ♘f7+ ♔g8 44 ♘e5 ♖d8 45 ♔g2 ♔f8 46 h4 ♗e8 47 ♔f3 ♗f7 48 ♔f4 ♔e8 49 ♖b1 ♔f8 50 b5 and, unable to prevent White from penetrating the b-file, Black rightly resigned. This was a simple example showing how to play across the whole board, thanks to White’s total control over the black position. We’ll look at more advanced examples in the section entitled Manoeuvring with Several Pieces, positions 103-108. 44 Teichmann – Chigorin Cambridge Springs 1904 Black to move The creative genius Chigorin played 14...♗xc3 15 bxc3 b5!. The exchange, followed by ...b5, secures Chigorin an outpost for his knight on d5. The chess world had to wait another 25 years until Nimzowitsch played in a similar way with pawns on b5 and f5, which shows how far Chigorin was ahead of his time. Grekov wrote the following about this position: “On looking at the apparently so solid position of White, who himself is ready to attack on the king’s flank, it is hard to believe that in only a dozen moves his king position, at present far removed from the influence of the black pieces, will be completely routed, the more so, since the player of the white pieces was a prominent master of defence. However, his opponent was an incomparable master of attack...” 16 ♖hg1 ♕e7! The d5-knight is so strong that the black kingside attack will decide very quickly. 17 ♖df1 ♕a3+ 18 ♔d2 b4 19 c4 ♗a4 20 ♕b1 ♘c3 21 ♕a1 ♖d8! An x-ray against White’s king. 22 g4 ♘e4+ 23 ♔e2 ♘c5! 24 ♕b1 ♘xd3 25 ♕xd3 ♕xa2+ 26 ♔f3 ♗c2 White resigned. The main reason White lost so quickly was that the centralised e5-bishop was unable to assist the defence. 45 Suba – Nicholson Malaga 2001 White to move Suba played the paradoxical 10 ♗xf6 ♗xf6 11 ♕d3. Why exchange pieces when White has a space advantage? Black’s Hedgehog position is an extremely harmonious opening if White tries to exploit his space advantage in the traditional way, since then Black’s counterplay can be tremendously strong. Black’s position is very flexible and his pieces look more passive than they really are – and for this reason it’s regarded as correct to exchange several minor pieces. In accordance with the same principle it has been discovered that it’s good to exchange queens, since then it will be easier to advance the pawns on the wings without worrying too much about dangerous counterattacks. The moral of the story is that one shouldn’t always avoid piece exchanges when having a space advantage, since everything depends on the concrete nuances of the position. Be aware of the fact that the Hedgehog is an exception to the rule. 11...♖a7 Black defends the bishop in order to avoid 11...0-0? 12 ♘g5 ♗xg5 13 ♗xb7. 12 ♖ad1 ♗e7 13 ♘d4 White exchanges another pair of minor pieces. 13...♗xg2 14 ♔xg2 ♕c8 15 f4 ♘c6 15...g6 was a safer move. 16 ♖f3 0-0 17 ♘xc6 ♕xc6 Three pairs of minor pieces have been exchanged so far. 18 f5 ♖c7 19 b3 In this position it’s clear that it is more difficult for Black to carry out the traditional Hedgehog breaks ...b5 or ...d5. White’s systematic and clever exchanges have led to Black’s reduced control of the white squares in his position. 19...♖b7 20 a4 ♖bb8 21 ♕e3 ♖be8 Nor is 21...♗f6 22 ♘e4 any improvement. 22 fxe6 ♗f6 23 ♕e4! ♕xe4 Black cannot move the queen without allowing ♖xf6 followed by e7, trapping the f8-rook, so the exchange is forced. 24 ♘xe4 ♖xe6 25 ♘xf6+ gxf6 The double rook ending is clearly advantageous for White since Black’s pawn structure is problematic with its four pawn islands. 26 ♔f2 ♖fe8 27 ♖d2 ♔f8 28 ♖fd3 ♔e7 29 ♖d5 ♖g8 30 a5 ♖b8 31 ♖2d3 ♖e5 32 ♖xd6 ♖xa5 33 ♖d7+ ♔e6 33...♔f8!. 34 ♖a7 ♖a2? The decisive mistake. Correct was 34...♖f5+ followed by 35...a5. 35 ♖e3+ ♔f5 36 ♖xf7 ♔g6 37 ♖ee7 and White won. 46 Hübner – Korchnoi Candidates match, Merano 1981 Black to move 10...♘xd4+! Normally piece exchanges should be avoided when playing with an isolani, but if you exchange the right pieces and are happy with a draw it can prove to be the right strategy. 11 ♘xd4 ♗c5 12 ♖d1 ♘e7 13 ♗e3 47 Black to move After the exchanges 13...♗xd4! 14 ♗xd7+ ♔xd7 15 ♖xd4 White’s position looks good, but the problem is that it’s difficult for the dark-squared bishop to make a difference with the d5-pawn located on a white square. Therefore White cannot easily improve his position and so Black’s drawing chances are almost 100 percent. 15...♖hc8 16 c3 16 ♔d3!?. 16...♖c4! 17 ♖ad1 ♖xd4 18 ♗xd4 f6 19 c4!? If 19 ♗c5 ♖e8! 20 ♔f1 ♔c6! 21 ♗xa7? b6. 19...♔c6 19...dxc4?! is met by 20 ♗xf6+ ♔e6 21 ♗xe7!? (21 ♗xg7 ♖g8 22 ♗d4 was the alternative.) 21...♔xe7 22 ♖d4 ♖c8 23 ♔d2. 20 cxd5+ ♘xd5 21 g4 ♖e8+ 22 ♔f3 In this position Korchnoi should have played 22...a6! and thereby increased his drawing chances. Instead he continued 22...b6. 48 Kramnik – Zviagintsev Tilburg 1998 White to move Here Kramnik played 27 ♗b5!. It might seem strange that White exchanges his “good bishop” for Black’s “bad bishop” but the reason is that Black’s pawns on white squares will be weaker after the exchange. Maybe it was this concept Larsen thought about when he once wrote that beginner’s books always lie when discussing the concept good and bad bishop. Very often the passive bishop, as Philidor already realised in the 18th century, is strong in defence. From that perspective it’s not at all strange to exchange such a defensive bishop. Kramnik said about his move: “My opponent underestimated the standard priyome ‘A bad bishop is also one that can defend many pawns.’ Once the bishops are gone, Black has to lose the a- or dpawn.” 49 Petrosian – Bannik Riga 1958 White to move Petrosian played the non-standard super-move 18 ♗c5!!. This is creative technique since it’s surprising that Petrosian exchanges his good bishop for Black’s bad bishop. He could adopt the standard plan and follow in Capablanca’s footsteps by playing 18 ♖xd8 ♖xd8 19 ♗xb6+ (19 ♖d1 ♖xd1+ 20 ♔xd1 ♘c8 avoids the exchange of the knight.) 19...♔xb6. The next step is to secure the f5-square with 20 g4, exchange the last rook with ♖d1 and then head for a position with a knight on f5 and the king on e4 after the manoeuvres ♘g3-f5 and ♔c2-d3-e4. We cannot reckon on achieving an ideal position with the king on f5 and the knight on e4 because Black will cover the f5-square with his own king on e6. A plausible continuation is 20...♔c7 21 ♖d1 ♖xd1+ 22 ♔xd1 ♔d7 23 ♘g3 ♔e6 24 ♔c2 ♗f8 25 ♔d3 ♗e7 26 ♘f5 ♗f8 27 ♔e4. White’s knight and king are stronger than their counterparts and he is in full control of the light squares – yet he cannot win this position. To realise this is to go beyond Capablanca’s technique. When we have understood the position is a draw we have to find another solution. This is called creative technique as compared to routine technique. 18...♖xd1+ 19 ♖xd1 ♗xc5 20 ♘xc5 ♖e8 21 ♘e4 It is now clear that it’s better to exchange the bishop than the knight since Black’s knight cannot cover both f6 and d6. 21...♖e6 21...♖f8 allows the invasion of the rook with 22 ♖d6 ♘d7 (or 22...f5 23 ♖f6!) 23 ♖e6!. 22 g4 White has a clear positional advantage and won after 61 moves by focusing on the weak light squares in Black’s position. 50 Taimanov – Bronstein Candidates tournament, Neuhausen/Zürich 1953 Black to move Bronstein played the correct exchange 18...♗xc3!. He didn’t want to grab the pawn by 18...♖xa2?, because of 19 ♖xa2 ♖xa2 20 e5! when Black has no targets on the queenside whereas White will have attacking chances in the centre. 19 bxc3 ♕a5 White has more than enough on his plate, since he now has to take care of three pawn weaknesses on a2, c3 and e4. 20 ♕d3 ♕a6 21 ♕d2 White should avoid the exchange of queens since then his weaknesses will be further accentuated. Variations that highlight this are 21 ♕xa6 ♖8xa6 22 ♖e2 (22 e5 ♘c7 and the d5-pawn becomes weak or 22 ♖ab1 ♖xa2 23 ♖b8 ♖a8 24 ♖eb1 ♖a1 and Black is in full control. White has problems with c3 and e4 while Black will have no trouble defending the e7-pawn.) 22...♘f6 and the e4-pawn cannot be protected without a2 falling. Nor can this pawn advance since this would leave d5 hanging. 21...♖xa2 22 ♖xa2 ♕xa2 23 e5 White shouldn’t have allowed the queen exchange. 23...♕xd2 24 ♘xd2 dxe5 25 ♖xe5 ♔f8 26 ♘b3 26 ♔f1 was a better opportunity to achieve a draw since the possibility of exchanging knights, in conjunction with sacrificing the pawns on c3 and d5 in exchange for the c5-pawn, would have resulted in a drawish position with four pawns versus three on the kingside. 26...c4 27 ♘c5 ♖a1+ 28 ♔h2 ♘f6! Now White realised that he couldn’t attack the e7-pawn since Black has prevented the knight check on d7 while preparing an attack on the c3-pawn with his rook. 29 ♘e4 ♘d7 30 ♖g5 ♖a2 31 ♖g4 f5 32 ♖f4 ♘b6 33 ♘g5 ♘xd5 34 ♖d4 34 ♖xc4 ♖xf2 35 ♘xh7+ ♔g7 36 ♘g5 ♖xg2+ 37 ♔xg2 ♘e3+ and Black wins the knight ending. 34...♘b6 35 ♖d8+ ♔g7 36 f4 h6 37 ♘e6+ ♔f7 38 ♘d4 ♘a4 39 ♖c8 ♘xc3 40 ♖xc4 ♘d5 41 ♘f3 ♖xg2+ 42 ♔h1 ♖f2 White resigned. 51 Gelfand – Malakhov Sochi 2005 White to move 14 ♗f4! White avoids the exchange of the dark-squared bishops for two reasons: a) White has a space advantage and should avoid the exchange, in accordance with Tarrasch’s famous rule. In this particular case it would only have helped Black’s queen find a good square on e7. White has a space advantage since the c4-pawn is more active than the c6-pawn and it’s easier for White to play e3-e4 than it is for Black to play ...e6-e5. b) It’s useless and planless to exchange since Black has no weaknesses and so it would then be easier for him to take advantage of his flexible structure. For example after 14 ♗xe7 ♕xe7 15 e4 dxc4 16 bxc4 c5 it is White who risks being saddled with pawn weaknesses. 14...f5 Black transposes to a Stonewall structure to prevent White from playing e3-e4, but his position loses flexibility. Gelfand writes insightfully in Positional Decision Making that in the Stonewall the g2-bishop is normally worse than a “bad” bishop on b7, a6 or h5. It’s not even certain that a bishop on b1 is better but it can be easier to exchange than when it is fianchettoed. 15 ♖fd1 ♗f6? The bishop has no function here apart from increasing Black’s control of the e5-square and making room for his queen on e7. According to Gelfand Black should take it easy and play 15...♖c8. One computer program suggests 15...g5 16 ♗g3 g4. 52 White to move And now the strange-looking move 17 ♘d2. With regard to this passive knight retreat Gelfand has written in the above-mentioned book that a friend of his had studied all of Fischer’s games and noticed that he never retreated his pieces. If his pieces were attacked he counterattacked rather than retreated. Gelfand was very impressed by this observation and thereafter always looked for counterattacks in similar situations. White could play the more active 17 ♘e5 ♘xe5 18 ♗xe5 ♗f6 19 ♗f4 and although White has agreed to one exchange he retains the easier position due to his superior king’s position. However, Komodo 11 doesn’t agree and suggests 17 ♘e1! dxc4 18 bxc4 ♕e8 19 ♘a2 ♕f7 20 ♘xb4 axb4 21 ♘d3 with a slight advantage. The knight has a very flexible and centralised position on d3 and can advance to e5 or f4 at the right moment. It’s actually more logical to exchange the strong b4-knight for the bad c3-knight and avoid the exchange of the good knight on e5 for the passive one on d7. 16 ♘a2! ♕e7 17 ♘xb4 ♕xb4 18 ♗d3 ♖fc8? Correct was 18...♕xb3 19 cxd5 ♗xd3 20 ♖xd3 ♕xd5 21 ♘e5 b5 22 f3 with a slight advantage. 19 cxd5 ♗xd3 20 ♕xd3 cxd5 21 ♖c2 ♕a3 53 White to move Although Gelfand eventually won the game, surprisingly he missed the opportunity of 22 ♗c7!, which is a well-known device to take over a file, on the same lines as the model game Karpov – Unzicker, which is presented as position 82 in the section on manoeuvres. White doubles rooks behind the protected bishop and in this clever way conquers the c-file and wins the strategical battle in a much more efficient manner. In the game the incorrect 22 h3? was played. A strong alternative to the Karpov-plan just mentioned was to establish a pawn wave on the kingside. White could increase his dark-square control with 22 g3! followed by ♔g2 and h2-h4. This kind of positional manoeuvre would consolidate and improve White’s game so much that he is then positionally winning. This idea of strong dark-square control is reminiscent of the famous game Schlechter – John, position 43. 54 Bronstein – Gaevsky Dnepropetrovsk 1939 White to move Bronstein played 9 ♗f1. At first glance it might seem strange to develop the bishop to b5 and then go back to f1, so what’s the point? The reason is that White has a space advantage, an e4-pawn versus a d6-pawn, and so Black wants to exchange as many minor pieces as possible to create more space for his pieces. By placing the bishop on f1 Black can only exchange one knight. This kind of move made by Bronstein is a standard choice which appears in many other positions to create a state of harmony. The bishop isn’t in the way of any other pieces on f1, since there is no other piece which needs that square for manoeuvre. If White had instead placed the bishop on e2 it would have been in the way of the rook, queen and the two knights and then we could say that White’s position lacked harmony. In the same way as a Mozart symphony, where no bars can be exchanged for any others, the bishop had to find the truly correct location and consider the prospects of the whole position and not only its own. After all, on e2 it would work in two directions, whereas now it only operates on the f1-a6 diagonal. As far as possible it is important to avoid putting pieces where they are in each other’s way. This is sometimes difficult to avoid in the opening phase when the board is more crowded, but in this particular position White has managed this very well. The only concession is that the f1-bishop works on one diagonal only but that is a small price to play for harmony. Compare this bishop with Black’s on e7 which for the moment cannot even move. However, eventually it will probably be located on f8 to create harmony and then be activated on g7. Alternatives to Bronstein’s move were either 9 ♗xc6 or 9 ♘xc6 with various pros and cons. It’s a matter of taste which you play but the main thing is to agree to the exchange of only one minor piece so as to be able to profit from the space advantage. The game continued: 9...♖e8 10 b3 ♗f8 11 ♗b2 g6 12 f3 ♗g7 13 ♘de2 White evades another exchange to avoid giving Black more space for his pieces. Black is a little cramped but his position is solid without any weaknesses. 55 Unzicker – Fischer Olympiad, Varna 1962 Black to move 16...♗b5! The concept to allow piece exchanges when it improves the pawn structure was something not only Philidor liked but also Fischer and Capablanca, two of the greatest transformation experts. 17 ♗xb5 axb5 18 ♘d5 ♘xd5 19 ♕xd5 19...♖a4! Here the same idea appears again. Black wants White to exchange on a4 and improve the pawn structure even further. Black could have exchanged on a1 or moved the rook to c8 since White has tactical weaknesses in his position on e4, c2 and b2 as well as a badly placed king and a passive knight on b3. To borrow an expression by GM Suba: the “bad” bishop on e7 protects a good pawn on d6. However, the bishop may have a more positive future since it can sometimes find a good location on g5 or manoeuvre to the classical diagonal on b6 via d8. 20 c3 ♕a6 21 h3? 21 ♖ad1 ♖c8 22.♘c1 gives Black a strong break with 22... b4. After 23 ♘d3 bxc3 24 bxc3 ♖a5 (24...♖xc3 was possible despite the fact that the last rank is weak. For example 25 ♘xe5 dxe5 26 ♕xe5 [26 ♕d8+ ♗f8] 26...♗b4 27 ♕xc3 ♕xf1+!) 25 ♕b3 ♖a3 26 ♕b1 ♖axc3 27 ♘b4 ♕a7 28 ♘d5 ♖3c6 29 ♕b3 and White had enough compensation for the pawn in Tal – Fischer, Curaçao 1962, which ended in a draw after 58 moves. Hübner has analysed 21 g3! with the plan of safeguarding the king on the third rank. After 21...♕a8 (or 21...♗g5 22 h4) 22 ♕xa8 ♖fxa8 23 ♔g2 followed by ♔f3 the game will most likely end in a draw. 21...♖c8 22 ♖fe1 h6 23 ♔h2 ♗g5 24 g3 Fischer suggests 24 ♖ad1 but Black has won the positional battle since the a-file is in Black’s control. 24...♕a7 25 ♔g2 ♖a2! When controlling a file, the goal is to take over the seventh rank as Nimzowitsch has noted. Black has managed this very well. 26 ♔f1? ♖xc3! White resigned. 56 Kramnik – Carlsen London 2012 Black to move 18...♗c6! This move, which Carlsen didn’t make, is reminiscent of 16...♗b5! which Fischer played to improve the pawn structure in accordance with Capablanca’s concept. But in what way does the pawn position improve, since a pawn on c6 is just as isolated as a pawn on b7? Incidentally, Carlsen played the less precise 18...♗xg2?! 19 ♔xg2 ♕d7 20 ♕b3 ♖fc8 21 ♗b2 ♕d6 22 ♖b1 ♖c5 23 ♖b6 ♖c6 24 ♖xb7 ♖xb7 25 ♕xb7 ♖b6 26 ♕c8+ ♕f8 27 ♕d7 h5! and managed to draw after very strong defensive play. Certainly not a bad feat in this kind of position against the mighty Kramnik. 19 ♗xc6 bxc6 A pawn on c6 is easier to defend than a pawn on b7. Black can place a rook behind the pawn, when it has more space compared with a rook behind the b7-pawn. White’s bishop cannot strike a rook on c8 as easily as one on b8. 57 Bogoljubow – Janowski New York 1924 Black to move Janowski played 19...♘e5! and sacrificed a pawn for the bishop pair. This was a concept Janowski liked and is good to know whether you play with or against the bishop pair. 20 ♗xh7+ 20 ♕h5?? f5 was obviously not playable and 20 ♕e2 ♗a4 21 ♕h5 ♘g6 would have retained the bishop pair, although Black’s bishops will become strong anyway since ...♗f6 is coming next move. 20...♔xh7 21 ♕h5+ ♔g8 22 ♕xe5 ♗f6 23 ♕h5 ♗a4 Black’s bishops have generated a lot of activity and Black has good compensation for the pawn. 24 ♖e1 ♕d6 25.h3 ♗c2 Komodo 10 calculates 14 moves ahead and comes up with 25...e5. Although Janowski’s move isn’t the most precise it is advisable to note this bishop’s incredible mobility. 26 ♕f3! b5 27 ♕e2 ♗a4 27...♗e4 was more precise. 28 ♕f3 28 ♗a1! was better to keep the bishop protected and avoid finesses such as ...♗c2 when White plays actively with the queen. 28...♖c4! Black plays like the exponents of the English school, founded by Staunton, and conquers the important key square c4. 29 ♗a1? ♖dc8? 29...e5 30 ♖xc4 (30 ♘f5 ♕d2 31 ♖b1 ♗c2) 30...bxc4 31 ♘e2 ♗d1! followed by ...♕d2 leads to a winning game. Note the a4bishop, which is the main reason Black’s advantage increases. 30 ♖b1 e5 31 ♘e2? 31 ♕g3! was correct. 31...♗c2 32 ♖bc1 ♗e4 The white-squared bishop is a monster on the long diagonal. 33 ♕g4 33...♗b7 Stronger was 33...♖xc1! 34 ♘xc1 (34 ♖xc1 ♖xc1+ 35 ♘xc1 ♕d2 36 ♘e2 ♕d1+ 37 ♔h2 ♗d3 and Black wins the knight or the bishop.) 34...♖c4. 34 ♖xc4?? 34 ♖ed1 is better. 34...♖xc4 The game is over since Black’s queen will now penetrate the white position. 35 f4 ♕d2 36 ♕g3 36...♖e4?! 36...exf4 37 ♘xf4 ♗xa1 38 ♖xa1 ♖xf4! 39 exf4 ♕d4+ would have left Black a piece up. 37 ♗c3 37 ♘c3 is met by 37…♖c4. 37...♕d5 37...♕d3 was stronger. 38 ♗xe5? 38 ♕f3!. 38...♖xe3! 39 ♕g4 39 ♕xe3 ♕xg2 mate. 39...♗xe5 40 fxe5 ♖xe5 41 ♔h2 ♕d2 41...♖g5? 42 ♘f4. 42 ♕g3 f6 43 h4 ♗d5 44 ♕f2 ♗c4 White resigned. What determined the strength of the bishop pair was the lightsquared bishop which decided the game. This was the main idea of the pawn sacrifice: to provoke White into exchanging his lightsquared bishop for the knight. Janowski is an interesting player to study if you want to focus on effective handling of the bishop pair – because Janowski simply loved the two bishops. Sometimes it’s difficult to decide whether it’s better to keep one rook or both rooks on the board and this decision often represents the difference between a win and a draw. Here are a few examples designed to unravel the interesting positional problem of when and why to exchange rook(s). 58 Vogt – Rukavina Olympiad, Skopje 1972 White to move One of White’s ideas to play for a win is to advance f5-f6. This plan will be significantly stronger if all rooks are swapped, since that would eliminate possible counterplay as well as creating tactical weaknesses in Black’s position. 26 ♖xd8+! ♖xd8 27 ♖d1 ♖xd1 28 ♔xd1 One might think that a four-rookswap benefits Black who is more cramped, but as a matter of fact it will then be easier for White to advance his f-pawn undisturbed. 28...♘d8 Note that 28...♕d7+ 29 ♔c2 ♘d4+ 30 ♗xd4 cxd4 is a bad version of a Capablanca ending with queen and knight against queen and bishop. 29 ♘d2 ♔d7 30 ♕e4 ♔e8 31 ♔e2 ♕a7 32 f5! ♕d7 33 f6! White carries out the winning breakthrough. 33...gxf6 34 exf6 ♗f8 35 ♕a8 ♗d6 36 g5! e5 36...hxg5 37 h6 is one concrete example why it’s good to swap the rooks. Black cannot defend his vulnerable last rank. 37 gxh6 ♕f5 38 ♕f3 ♕h7 39 ♕d5 ♗c7 40 ♗xe5 Black resigned. In this game it became easier for White (and Black) to play with just queen and king but White benefited the most due to the decisive pawn break f5-f6. 59 Vidmar – Nimzowitsch New York 1927 Black to move Many players would probably have moved one of the rooks to d8 and thereby removed the inherent dynamic in the position after all the exchanges on the d-file had taken place. However, instead Nimzowitsch played the profound: 19...♖ae8!. Black must exploit his qualitative pawns on e5, f5 and g5 by placing the major pieces behind these pawns. White does not have any dangerous square of invasion on the d-file. White’s rook can come to d6 but it doesn’t mean anything since there are no weaknesses on the sixth rank. 19...♕g7!? was an interesting alternative and then later place the queenside rook on e8. 20 ♕a4 ♗a8! Black must parry the threat of ♖d7. 21 ♖d6 ♕g7! 22 ♗f1? It was better to play 22 ♗e1 g4 (22...e4? 23 ♗c3) 23 fxg4 ♘xg4 24 ♗xg4 ♕xg4 25 ♕c2. 22...e4 23 ♗e1 (23 fxe4 ♘xe4) 23...exf3 24 ♗c3 ♕e7! 25 ♖6d3 fxg2 26 ♗xg2 ♗xg2 27 ♗xf6 (27 ♔xg2 ♕e4+) 27...♕e4! 28 ♖1d2 ♗h3 29 ♗c3 ♕g4+ White resigned. This was creatively played by Nimzowitsch, who realised, contrary to the classical school, that it is not necessary to oppose an enemy rook on an open file, as long as the file cannot be exploited in any way. 60 Stoltz – Gligorić Saltsjöbaden 1952 Black to move 10...♗d7! When White has a structural weakness, like here on d4 where Black wants to place his knight, it’s advisable for Black to avoid the exchange of rooks. This is because it’s easier to realise the advantage of a strong outpost if there are rooks on the board. Moreover White’s rooks will become passive as a consequence of the strong centralised knight. 11 ♘d5 This outpost is of a temporary nature compared to Black’s on d4. 11...h6 12 ♗e3 12 ♘xf6+ exf6 gives Black the possibility to break in the centre with ...f6-f5. 12...b6 13 ♘ec3 ♘e8! 14 ♗c1 e6 15 ♘e3 ♘c7 15...♘d4 is too early due to 16 ♘e2. 16 ♗d3 ♘a6 17 f4 17 a3 weakens b3 which can be exploited by a knight on d4. 17...f5 18 exf5 18 e5 g5 is advantageous for Black. 18...exf5 19 ♘cd5 ♘ab4 20 ♗b1 20...♗d4 One advantage of waiting with the occupation of d4 is that you retain the option of placing another piece there. 21 a3 ♘xd5 22 cxd5 22 ♘xd5 ♘a5 23 ♗a2 ♗a4 makes the b3 weakness palpable. 22...♘e7 23 ♗a2 ♔g7 24 ♘c2 ♗f6 25 ♔f2 ♗a4 26 b3 ♗e8 27 ♖he1 ♗f7 28 b4 ♘xd5 29 bxc5 bxc5 30 ♖d3? ♘xf4 White resigned. In contrast to his contemporaries, Morphy realised that when you are ahead in material it makes sense to simplify by exchanging queens. First of all you eliminate counter-chances and then secondly you approach an endgame. We will not look at such obvious cases but instead examine other perspectives why an exchange of queens might be useful. 61 Mnatsakanian – Petrosian Bundesliga 1993 White to move 13...♕f6! It’s logical to strive for an exchange of queens, so as to take the sting out of White’s advantage in space. Also it wasn’t clear what else the queen might do. It’s a clever idea to develop the bishop to d6 (instead of e7) and make possible the queen manoeuvre to f6 challenging White’s active queen. In the game play continued 14 ♕xf6 ♘xf6 after which White’s centre was automatically neutralised. Petrosian won after only 31 moves. The other option 14 ♗xd5 cxd5 15 ♕xf6 gxf6 offers no tangible advantage for White because it’s difficult to exploit the disrupted pawns on the kingside if Black keeps his king in that area. Instead, 14 ♕e4 may have led to a sharper game. We don’t know on which side Petrosian might then have castled. In similar typical structures he used both options. 62 Rozentalis – Appel Moscow 1964 Black to move 29 ♕a3! A deep concept. White effectively sacrifices at least half a pawn, since the b-pawn will be placed on a3 behind the a5-pawn, but in return White gains control of the b-file. Along that file White can not only exert pressure on b7 but also on d5 via the b5-square. 29...♕xa3? 29...♖d8 would have upset White’s plans more. Now he gets what he wants. 30 bxa3 ♖d8 31 ♖b2 ♖c7 32 ♖b5 ♖dd7 33 ♔f2 g6 34 ♔e3 ♔g7 35 ♖fb1 ♔f7 36 ♖c5! ♔e7 36...♖xc5 37 dxc5 and White’s king controls the d4-square. 37 ♖bb5 ♖xc5 38 dxc5! ♔d8 39 a6 ♔c8 40 ♖b6!! It’s amazing what moves the major pieces are capable of! 40...♗g8 41 ♖f6 ♖d8 42 ♔d4 bxa6 43 ♖d6 Black resigned. An extraordinary display which to my eyes is Rozentalis’s immortal game. 63 Smyslov – Tal Candidates Tournament, Yugoslavia 1959 White to move 15 ♕d3! White benefits from an exchange of queens, for which the doubled pawn on the d-file is a small price to pay. The reason is that White can then start play on the c-file and together with the bishop pair and the strong a5-pawn this will ensure at least a slight advantage. Note that White also prepares c2-c4 if Black should avoid the exchange. 15...♖fc8 16 ♖fc1 ♕xd3 16...♕c4 17 ♘d2. 17 cxd3 g6 17...♘c5? 18 ♘xc5 dxc5 19 d6! wins material. 18 ♖c3 ♖xc3 White was threatening to double the rooks but this exchange helps to improve White’s pawn position and open the b-file. 19 bxc3 ♖c8 20 c4 20...e4! Black must play actively, otherwise White places his rook on b1 and creates threats on b7. 21 dxe4 ♖xc4 22 ♘d2 ♖c2 23 ♗d1 ♖c3 24 ♔f1 ♘c5 25 ♗d4 ♖d3 26 ♗xc5 dxc5 If 26...♖xd2 27 ♗e3 ♖b2 28 ♖c1 ♘xe4 29 ♖c8+ ♔g7 30 ♗d4+. 27 ♔e2 ♖xd2+ 27...♖d4 is answered by 28 f3 and 27...c4 by 28 ♖a4. 28 ♔xd2 ♘xe4+ Black tries to construct a fortress with the help of strong darksquare control. Now Smyslov should have played 29 ♔d3! with a clear advantage in the endgame. 64 Spassky – Petrosian World Championship match, Moscow 1966 White to move 13 h4!! This very concrete move is a reaction to Black’s simple plan of ...♖ab8 followed by ...c6-c5. Other ideas like 13 g4 followed by h4 to strive for an attack and open lines on the kingside or 13 ♗d3 as well as 13 ♗c4 followed by ♖he1 are not so effective mini-plans. Spassky’s plan is to activate the rook on h3 and place the queen on g5 and then exchange queens so as to be able to exploit his better pawn structure and his bishop versus knight in the endgame. 13...♖ab8 14 ♖h3! c5 14...♖b6 is answered by 15 ♕g5! ♖fb8 (15...♕xf2?? 16 ♖f3 ♕g1 17 ♕g3 and Black’s queen is trapped.) 16 ♗c4 with the point that 16...♖xb2?? fails to 17 ♗b3 trapping Black’s rook. 15 ♕g5! This is the real point of Spassky’s clever 13th move. White isn’t interested in a kingside attack and prefers to exchange queens so as to play against the weak pawns with ♖c3 or ♖a3. 13 h4!! wasn’t preparation for an attack but rather preparation for an endgame! Ingenious play by Spassky! The game concluded quickly after the following moves: 15...cxd4 16 ♕xf6 ♘xf6 17 ♖xd4 ♖b7 18 ♖b3! ♖xb3 19 axb3 ♖a8 20 ♖c4 ♘e8 21 ♖a4 ♘d6 22 g3 ♔f8 23 ♗g2 ♖c8 24 ♖xa7 ♔e7 25 ♔d2 h6 26 c4 g5 27 hxg5 hxg5 28 ♔c3 ♔d7 29 b4 ♖h8 30 b5 ♖h2 31 ♗c6+ ♔d8 when Black resigned. 65-108: Manoeuvring All pieces have a specific task and this is most clearly seen in the manoeuvre. It is particularly in closed or locked positions that manoeuvring skill becomes most important. Other situations are when one of the players has the initiative or more space, since then it’s easier to regroup one’s forces more efficiently. First the knight’s capabilities are shown, then the capabilities of the bishop, rook and queen. Two examples of king manoeuvring can be found in positions 87 and 88. Another aspect of manoeuvring involves several pieces where one important concept is the principle of exploiting two weaknesses and playing over the whole board. 65-78: The Knight Manoeuvre With its exotic way of moving, the knight is the most magical piece on the chessboard and most of us have problems dealing with it optimally. Bronstein wrote that we usually devote our main attention to the long-range pieces (the queen, rooks and bishops) and discover too late the danger of pawns and knights when they have already crossed the fifth rank. “What evidently operates here is the same law as in life, which forces us when crossing the street to keep a constant look out for a speeding cyclist, in order to avoid having an accident.” Such an “accident” has happened to me more than once and a typical example is the following: 65 Engqvist – Matros Stockholm 1998 Black to move Black crossed the “equator”, to use Bronstein’s terminology, with 16...♘b4! and I played 17 a3 expecting something like 17...♘a6, but instead the knight dug deeper with 17...♘a2!!. This was obviously a shock to me so I replied 18 ♖c2? My opponent Matros expected me to play 18 ♖a1 and analysed 18...♗xc4 (18...b5 might be more precise.) 19 ♗d4 b5 20 ♗xg7 ♔xg7 21 ♗xa8 ♕xa8 22 ♖g1 bxa4? (22...f6 or 22...♔g8 was correct with excellent winning chances.) 23 ♕d4+ ♘e5 24 f4 with counterplay. 18 ♖b1 might have been the best reply but Black still takes over the game after 18...♗xc4 19 b3 ♗e6 20 ♕c2 b5. 18...b5! 19 ♗xa8? 19 ♘c3 was better. 19...♕xa8 20 f3 bxa4 21 ♕a1 21...♘b4! This knight is unbelievable! 22 axb4 axb4 23 ♕d1 b3 24 ♖d2 ♘e5 And here comes the other knight – and very fast! 25 ♗d4 ♘xc4 A second knight crosses the equator as well. 26 ♗xg7 ♔xg7 27 ♖d4 ♘xb2 This knight is also placed on the seventh rank. 28 ♕c1 a3 29 ♖h4 and White resigned because of 29...♕a5+ 30 ♔f2 ♕c5+. The moral of this “accident” is that it’s vital to observe the knight’s movements as early as possible, otherwise it may be too late... 66 Benko – Stein Interzonal, Amsterdam 1964 Black to move A simple form of manoeuvring is when a player seeks to improve the position of a certain piece in a few moves. Stein played 35...♘e8. The knight is better placed on d6 where it not only puts pressure on the e4-pawn and the c4-square but also prepares a future ...f7-f5 at a suitable moment. One might also say that this manoeuvre is in accordance with Philidor’s principle which states that the mobility of the pawns is the most important factor. This principle is normally valid in closed positions but sometimes also in semi-closed positions like here. 36 ♗g2 ♘d6 37 ♗h3 ♕a2+ 38 ♕c2?? 38 ♔f3 would have kept the balance. 38...♘xe4! It is always important to have a good eye for such tactical shots, especially in time pressure. Stein won the game easily after 62 moves. 67 P. Carlsson – E. Agrest Sweden 2016 Black to move 12...♘a7 Black plans the seemingly slow manoeuvre ...♘a7-b5-c7-d5 but it works, since the game has a closed character and the value of a lead in development decreases. The game continued 13 ♕e2 ♘b5 14 ♖d1 ♘c7 15 ♘fd2 when the best was 15...♘d5 after which Black has equalised thanks to the well-placed centralised knight. In the game the inferior 15...b5?! 16 axb6 ♘xb6 17 ♘a5 ♕d7 18 ♘db3 was played with a slight advantage to White. But Agrest managed to win the game anyway after 38 moves. 68 J. Polgar – D. Wood Hastings 1988 White to move Judit Polgar played 24 ♘b1!!. Black has a big hole on e5 so the knight manoeuvre ♘b1-d2-f3-e5 isn’t so difficult to find. You need to be familiar with such patterns as well as calculating four moves ahead. Note the really nice preparatory move ♔a2 which vacated the b1-square for the knight. 24...♕f6 25 ♕b4 b6 26 ♘d2 ♖ff7 27 ♘f3 ♖e4 28 ♕d2 ♖fe7 29 ♘e5 Now that the knight is placed on e5 all Black’s counterchances are completely gone. 29...♗b5 30 ♔b1 ♖4xe5? A natural reaction which shows the strength of the knight but of course doesn’t solve Black’s problems. It would be better to play a waiting game with 30...♗e8 but of course that is easier said than done in the heat of the battle. 31 fxe5 ♕xe5 32 ♖d1 h6 33 ♖gg1 ♕f6 34 ♕f4 ♖e4? 35 ♕xc7 Black resigned. In Karpov’s most famous game from the Candidates in 1974, against Spassky, Karpov played a similar manoeuvre ♘b1-d2-f3. I’m pretty sure Polgar was already familiar with the idea due to that game. You should look it up if you haven’t already seen it. 69 Fischer – Mobley Simultaneous Exhibition, USA 1964 White to move Compared with the previous examples 18 ♘e2! represents a slightly more complicated manoeuvre. First of all White manoeuvres his worst placed piece but the dream or long-term plan is to manoeuvre to either the e5 or c5 square. Let’s see how Fischer managed this because for the moment it’s not clear how it can be done. The game continued 18...♘b4 19 ♔b1 ♘xd3 20 ♖xd3 f4 21 ♖dd1 b4 22 ♘c1! Now the picture is clearer. Fischer wants to place the knight on b3 or d3. 22...a5 23 ♕e2 ♖f6 24 ♕e5 ♕d8 25 b3 ♖b5 26 ♕e2 ♕a8 27 ♕c4 ♗f5 The moment has come for our hero. 28 ♘d3! ♖d5? Black should have eliminated the knight with 28...♗xd3. Now the strongest move was 29 ♘e5! with a clear advantage to White. Instead Fischer played 29 ♘xf4 and won quickly after 29...♗xc2+ 30 ♕xc2 ♖xf4 31 ♖xe6 ♖fxd4 32 ♖xd4 ♖xd4 33 ♖xc6 ♕d8? 34 ♖c8 ♖d1+ 35 ♔b2 a4 36 ♖xd8+ ♖xd8 37 bxa4 Black resigned. 70 Kramnik – Grischuk Candidates tournament, Berlin 2018 White to move The knight isn’t doing anything on f3 so where should it manoeuvre? 18 ♘g5!? A more conservative move was 18 ♘e1 but Kramnik plans the mini-manoeuvre ♘h3-f4 with an eye on d5. 18...♖c5 If Black tries to cut off the knight from the central squares with 18...h6 19 ♘h3 g5?! White responds 20 f3 followed by ♘f2 and then he can enjoy the weakened state of Black’s kingside. The most critical move however was 18...b5!, exploiting the fact that if White takes twice on b5 Black sets up a double threat on g5 and b5 by ...♖c5. 19 ♘h3 b5 20 ♘f4 bxc4 21 ♗xc4 ♖g5!? Grischuk has been creative and manoeuvred his rook to g5 where it wants to provoke weaknesses on the kingside. However, here it risks becoming cut off. A safer move was 21...♘e5. 22 ♘cd5 Note how active and easy to play White’s position becomes. The battery on the long dark-squared diagonal is opened, exerting positional pressure on the black position. 22...♘e5 23 ♗e2 ♘e4? 23...♗c6 with the idea ...♕b7 was correct. 24 ♖d4 ♘c5 25 h4 ♖f5 26 e4 The computer recommends 26 g4! with a clear advantage but that move introduces glaring weaknesses on the kingside and is apparently not for humans. 26...♖xf4 27 ♘xf4 ♘xe4 28 ♘d5 Kramnik managed to exploit his small material advantage and won after 48 moves in a superbly played game. 71 Lékó – Kramnik Budapest 2001 Black to move Kramnik is especially skilful when it comes to knight manoeuvres so let’s see one more example. Here he played 22...♘e7!!. What’s going on, since the e4bishop dominates the knight? During the press conference Kramnik said that he wanted to be active on the kingside. And what activation! This is an instructive example showing how much kinetic energy a knight really can possess when the circumstances are right. Bronstein once wrote that “what is difficult is to see a knight from afar, when this piece is only preparing for a deep raid into your defensive lines.” We are actually missing the fact that the knight is a long-range piece in slow motion. 23 ♖e2 f5 This is the point. 24 ♗f3 24 ♗d3 obviously allows the strong activation 24...♘d5. 24...g5 25 fxg5 hxg5 26 h3 ♔f7 27 ♔e1 ♔f6 28 ♖d2 ♔e5 29 ♔d1 ♘g6 The knight is on its way... 30 g3? White creates an unnecessary weakness on the wing where Black aims his attack. 30...♖h8 31 ♗g2 f4 32 ♖d3 fxg3? Even stronger was actually to return with 32...♘e7!. 33 ♖xg3 ♔f5 34 ♔d2 ♘e5 35 ♗f1? 35 ♗d5 was correct. 35...♔f4 36 ♖b3 ♘f3+ The equator has been crossed with devastating effect. If one compares this position with that in the initial diagram then it can be seen that an incredible metamorphosis has taken place. It’s rare to see a knight manoeuvre like this and one that is so decisive. 37 ♔e2 g4! 38 hxg4 ♖h2+ With the help of the pawn break Black obtained good winning chances and Kramnik won after 133 moves. 72 Steinitz – Lasker World Championship match, New York 1894 White to move Steinitz played the retreating knight manoeuvre: 18 ♘h3! ♘c5 19 ♘f2 According to Steinitz’s theories a flank attack can only succeed with a secure centre. Note that the Spanish bishop was previously placed on b3, exerting pressure on f7, and that the knight was placed on g5 in an attacking position. Now these two pieces have the defensive functions of securing the central structure, in particular the d3-square. The flexibility of how to use the pieces, the oscillation between attacking and defending, is typical of Steinitz. 19...b5 20 g5 h5 21 ♘f5! Now there is a switch from defence to attack. Another passive knight suddenly springs to life as well as the c1-bishop. 21...gxf5 22 exf5 f6 23 g6 ♘xg6 24 fxg6 ♗xg6 25 ♖g1 and Steinitz had a strong attack resulting in a win after 42 moves. The process of manoeuvring a knight to g5 and then back to h3 is also a modern idea explored in detail by GM Adrian Michalchisin on one of his DVDs about making plans. Among other modern players, Korchnoi has been an exponent of this interesting idea. 73 Engqvist – Heinola Stockholm 2008 White to move According to a famous Russian rule, if it’s not possible to find an appropriate plan one should improve the worst placed piece – because this is an effective mini-plan in itself. But what to do if you are saddled with two passive pieces as in this position? White’s problem pieces are the rook on a1 and the knight on b3. Which of these pieces ought to be moved? A classical principle is to move the minor piece before the major piece but of course this must be checked when making calculations. In the game I didn’t play the best move. I continued with 24 ♖ad1?! when my opponent responded 24...♗g4. While I was thinking whether to play the weakening 25 f3 or the tempo losing 25 ♖a1 my opponent’s cell phone rang so I won the game immediately, in accordance with the strict rules we had for this occasion. I was lucky because Black’s position is to be preferred. The rook move to d1 invited dynamics but these were only in Black’s favour as his bishops are strong, particularly the one on the a7-g1 diagonal. Correct was 24 ♘e2! Another solution to the problem was 24 ♘c1! ♗g4 25 ♘ce2 which I missed during the game. 24...♗g4 25 ♘bd4 For some inexplicable reason I completely overlooked this possibility of manoeuvring one of the knights to e2. White’s position is advantageous from a static perspective and the knight manoeuvre is in accordance with the static nature of the position. It’s sometimes easy to make routine moves, bringing the rook from the left hand corner into the centre, but on this occasion it wasn’t the best option. What is correct is to manoeuvre the wrongly placed knight on b3 to e2 and this could have been achieved in two different ways! Knights are peculiar pieces indeed and we must always treat them in the way they deserve. This is also the main reason we have several examples with knight manoeuvres. 74 Colle – Capablanca Carlsbad 1929 Black to move What to do with the threatened knight when there are several options? Capablanca’s solution was... 13...♘c6!. Of course Black should pressurise the strong d4-knight. It’s stronger than the bishop on c1. 13...♘d3? is a positional mistake due to 14 ♖f3 ♘xc1 15 ♖xc1 and White is happy as he no longer has to worry about the problem bishop on c1. 14 ♖f3 g6! This move explains one of the reasons Capablanca didn’t want to place the knight on g6 at move 13. Black is playing in a prophylactic manner against White’s planned e4-e5. 14...♘xd4 15 cxd4 ♗e2 16 ♕xe2 ♕xc2 17 ♖c3 ♕a4 was another option. 15 ♘2b3 15 ♘xc6 ♕xc6 16 f5 ♗c5+ 17 ♔h1 ♘g4 is in Black’s favour. 15 f5 gxf5 16 exf5 e5 17 ♘xc6 ♕xc6 followed by ...♔h8 and ...♖g8 gives Black strong pressure on the g-file. 15...♘xd4 15...d5 was also good. 16 ♘xd4 ♗b7! 17 ♕e2 ♗c5! A move with multiple ideas. It stops the c1-bishop from coming out and it plans ...♕c6! with strong pressure on the classical diagonal as well as the long diagonal. 18 ♖h3? 18 ♖e3 with overprotection of the e4-pawn was necessary! 18...♕c6! …and Black won after only 34 moves. After the game Colle uttered the legendary words: “It seems to me I did not make a weak move, but that my opponent played some very strong ones.” 75 Fischer – Spassky Sveti Stefan 1992 White to move 27 ♘f1!! The beginning of an extraordinary knight manoeuvre to a3 via d2 and b1. Another way of tackling the position was 27 axb5 axb5 28 ♗a7 (Karpov’s ingenious idea which we will discuss later in more detail.) followed by ♖a5, ♘d2-b1-a3 and the b5-pawn will be difficult to defend in view of the black army’s limited room for manoeuvre. 27...♗e7 28 ♘1d2 ♔g7 29 ♘b1 You often see a Steinitz manoeuvre from b1-d2-f1-g3 but here Fischer does this the other way around, a reversed Steinitz manoeuvre, and that is highly unusual and original! Spassky was an active player who didn’t like passive positions and that is why the slightly desperate 29...♘xe4!? followed. You will see more such psychological moves by Spassky when looking at positions 148-149. Spassky wants to sacrifice one of his knights in exchange for smashing White’s strong pawn centre but of course it isn’t correct since White still has his positional pressure. 30 ♗xe4 f5?! A better continuation was 30...♘f6 31 ♘fd2 ♘xe4 32 ♘xe4 ♗xd5. 31 ♗c2 ♗xd5 32 axb5 axb5 33 ♖a7 ♔f6? It’s better to manoeuvre the king to the last rank and allow the h7-knight to use this square. 34 ♘bd2 ♖xa7 35 ♖xa7 ♖a8 36 g4! The king hunt has begun! 36...hxg4 37 hxg4 ♖xa7 Nor does 37...f4 help after 38 ♗e4! fxe3 39 ♗xd5 ♖xa7 40 ♘e4+ and neither does 37...fxg4 due to 38 ♘h2. 38 ♕xa7 f4 39 ♗xf4! exf4 40 ♘h4!! 40 ♕d4+ ♔e6 41 ♗f5+ also won but Fischer’s choice is more aesthetic. 40...♗f7 If 40...♘f8 then 41 ♕d4+ ♔e6 42 ♘xg6 ♘xg6 43 ♗f5+ ♔f7 44 ♕xd5+. 41 ♕d4+ ♔e6 41...♔g5 is answered either by 42 ♘g2 or 42 ♕g7. 42 ♘f5! 42...♗f8 Or 42...gxf5 43 ♗xf5 mate, 42...♗f6 43 ♕xd6 mate, and 42...♔d7 43 ♕a7+ ♔e6 44 ♘g7+. 43 ♕xf4 ♔d7 Or 43...gxf5 44 ♗xf5+. 44 ♘d4 ♕e1+ 45 ♔g2 ♗d5+ 46 ♗e4 ♗xe4+ 47 ♘xe4 ♗e7 48 ♘xb5 ♘f8 49 ♘bxd6 ♘e6 50 ♕e5 and Spassky resigned. One of Fischer’s immortal games! 76 Engqvist – J. Eriksson Stockholm 2017 White to move In this position I unfortunately missed the nice knight manoeuvre: 19 ♘b5! This concept of a double knight attack on the fifth rank is inspired by the maxim that one should always analyse aggressive variations which lead to a forced entry to the enemy position. This is Fischer’s well-known principle which states that you should only go forward with the pieces, not backwards. With the aggressive knight move White exploits the fact that Black hasn’t yet had time to play one of the moves ...a6 or ...c6 to secure the b5-square. 19...♘e6 19...♗xb5 20 cxb5 leads to a clear positional advantage since f2f4 will come with a space advantage and more active play. Eventually the c5-knight will be driven away as well after a timely b3-b4. 19...c6? is effectively answered by 20 ♘xd6 ♖e6 (20...cxd5 21 ♕xd5+ would lead to a substantial material advantage since the c5knight is hanging. If Black places the knight on e6 White captures first on e8 and then on e5.) 21 ♘f4 White has won a pawn for nothing. This variation is the most important and most difficult to foresee since it involves tactical elements. 20 f4 ♘g4 21 ♗xg7 ♔xg7 22 e4 and White has a clear positional advantage. Remember that knights should move forward in accordance with Fischer’s formula mentioned in position 52! 77 Pomar – Alekhine Madrid 1945 White to move Pomar retreated with 19 ♘b3 and that retreat triggered Alekhine to give his student Pomar one really good piece of advice. He said: “There are positions where a combination is obligatory.” The world champion recommended instead 19 b4 and if 19...bxc5 (Alekhine answered 19...♔b8 with 20 ♖eb1 ♔a8 21 ♘a6 c6 with adequate defensive resources but then it would have been more difficult to attack on the kingside.) 20 bxc5 ♕c6 21 ♖ab1 followed by ♗b5 with a very dangerous attack. Komodo 11 doesn’t agree and mentions 21...a6 with a slight advantage to Black. This correction doesn’t mean that Alekhine’s pedagogic piece of advice was useless, on the contrary. Komodo 11 actually recommends 19 ♘a6 with a slight advantage to White, proving that Alekhine was right after all with his insight that the piece should not have retreated. After the passive move in the game White’s position lost its sting and following 19...g5 20 ♘h2 ♘e6 21 c4 dxc4 22 ♗xc4 ♔b8 23 ♖ac1 ♘d8 24 ♗xf7 ♘xf7 25 ♕c4 ♘d5 Pomar was outmanoeuvred step by step and Alekhine won after 48 moves. Alekhine wanted to tell his student that White had to make an important decision: whether to sacrifice the knight with 19 b4, since that was the only way to keep the momentum (energy) in the position, or at worst move the knight. But according to Komodo it was correct to go forward with the knight in the spirit of Fischer. 78 Rubinstein – Tarrasch Carlsbad 1923 White to move Rubinstein was a great master of playing against an isolated pawn and he introduced important ideas relating to the appropriate technique. In this classical position he played 11 ♘e1!?. White plans the manoeuvre of the knight to d3 and then to f4 where it pressurises the d5-pawn. It’s not at all mandatory to blockade the isolani with a knight on d4 as Nimzowitsch recommended. A serious alternative option is to follow in Rubinstein’s footsteps and play around the isolani and perhaps even capture the pawn. The game continued 11...d4 12 ♗xf6 ♕xf6 13 ♘e4 ♕e7 14 ♘xc5 ♕xc5 15 ♘d3 ♕b6 16 ♘f4! ♕xb2? 16...♖fe8 was better. 17 ♘xe6 fxe6 18 ♖b1 ♕xa2 19 ♖xb7 and after 59 moves Rubinstein’s opponent resigned. This ends the section on knight manoeuvres. We now move on to bishop manoeuvres. 79-87: The Bishop Manoeuvre Bronstein used to say that “the queen’s bishop will never find a better target than the black g7-pawn.” So here we have a strong argument for 1 b3 if it were not for 1...e5 blocking the diagonal. To find a good diagonal for the bishop, where it can achieve maximum power, is always a matter of adapting to the pawn structure. There is no reason to play, for example, 1 d4 ♘f6 2 ♗f4 if Black can reply 2...d6, literally killing the bishop. So, better to be flexible and play 2 ♘f3 and if 2...d5 then 3 ♗f4. The first three positions show how the future prospects for a bishop may be improved by a paradoxical retreat of this piece to its starting square. In this way it could well find a more effective diagonal. The next three positions show how to take over a file by placing a bishop on the seventh rank in front of a rook. The last three positions handle a manoeuvre from the classical diagonal to the b1-h7 diagonal, bishops of opposite colour and the bishop pair in a symmetrical position. 79 Andersson – Browne Naestved 1985 White to move A bishop manoeuvre to e3 or f4 is a typical idea in the Hedgehog and therefore Andersson played 16 ♗c1! moving back to its original c1-h6 diagonal. 16...♘fe4 17 ♗f4 ♖c8 18 ♘xe4 ♗xe4 19 ♗d6?! Correct was 19 ♖d2 but if White prefers to manoeuvre his bishop a better try was 19 ♗e5, provoking 19...f6, and only then 20 ♗d6 since the f6-square is not available for the bishop. The exchange of the dark-squared bishops makes it easier for White to exploit the pawn majority on the queenside. 19...♗f6! 20 ♘e5 ♘b7 21 f4 ♘xd6 22 ♖xd6 ♗e7 23 ♖dd1 ♗xg2 24 ♔xg2 f6 25 ♘f3 ♗c5 26 ♖d3 ♖d8 27 ♖ad1 Black’s bishop hampers White’s attempts to mobilise his pawn majority so a draw was agreed. 80 Smyslov – Langeweg Amsterdam 1981 White to move 30 ♗c1 White plans b2-b3 followed by ♗b2 with pressure against the f6pawn. This must be a very strong move and you cannot help thinking of Bronstein’s statement: “A bishop, standing on a good diagonal aimed at the opponent’s king, is better than any rook or knight.” Surprisingly the computer prefers a manoeuvre with the other bishop. After the strong 30 ♗e2! White plans ♗g4 with further pressure on f6. 30...♘f7 (30...h5 31 g4!) is answered by 31 ♗g4 ♘fe5 32 ♗e6. White’s more active pieces decide the game with the main threat being the capture on e5. 30...♘f7 31 b3 ♖ab8 32 ♗b2 ♘fe5 33 a4 a6 34 ♗e2 ♔f7 34...b5 35 axb5 axb5 36 ♘b4 ♗a8 37 ♘cd5 and White has a winning positional advantage. 35 g4 ♔g7 36 a5 ♖d8 37 ♖f2 ♖bc8 38 ♖df1 Black lost on time, but his game was positionally lost anyway, since it’s not possible to save the f6-pawn. 81 Kasparov – Petrosian Tilburg 1981 White to move 11 ♗f1! A paradoxical retreat which shows a deep understanding of the requirements of the position. White wants to defend the h3-pawn so it will be possible to castle on the kingside. The bishop will be well placed on g2, observing the long white diagonal, and it will make it more difficult for Black to attack White’s central pawn with ...c5. This clever bishop move helps White to maintain his strong position in the centre. Kasparov mentions in his annotations that when he stood up from the board after this excellent move, Spassky slapped him on the shoulder saying: “A good manoeuvre!” It’s noticeable that Petrosian had this position twice with the white pieces but he didn’t play this clever preventive bishop manoeuvre. But, then again, we should remember that Petrosian was famous for his own profound manoeuvring skills. Instead he preferred the space-gaining move 10 g5. 11...c6 Black needs to play a passive move with his c-pawn to neutralise the long white diagonal. 12 ♗g2 ♕c7 13 0-0 ♗e7 14 f4 ♘b6 15 g5 White has a small edge but Black’s position is very solid. Petrosian eventually managed to win the game, which in fact became historic because it featured one of the most famous king moves ever played. 82 Karpov – Unzicker Olympiad, Nice 1974 White to move This position arose after the moves 1 e4 e5 2 ♘f3 ♘c6 3. ♗b5 a6 4 ♗a4 ♘f6 5 0-0 ♗e7 6 ♖e1 b5 7 ♗b3 d6 8 c3 0-0 9 h3 ♘a5 10 ♗c2 c5 11 d4 ♕c7 12 ♘bd2 ♘c6 13 d5 ♘d8 14 a4 ♖b8 15 axb5 axb5 16 b4 ♘b7 17 ♘f1 ♗d7 18 ♗e3 ♖a8 19 ♕d2 ♖fc8 20 ♗d3 g6 21 ♘g3 ♗f8 22 ♖a2 c4 23 ♗b1 ♕d8. Here Karpov played the stunning 24 ♗a7!! – a revolutionary idea which avoids any exchanges of major pieces. Often in the Spanish Game White plays on the kingside while Black gets counterplay on the other flank. But after the original move played in the game Black’s counterplay on the queenside is over. It is an important idea to remember whenever it’s possible to control a file by doubling or tripling. 24...♘e8 25 ♗c2 ♘c7 26 ♖ea1 ♕e7 27 ♗b1 ♗e8 28 ♘e2 White slowly prepares play on the kingside with f2-f4. Black cannot do much and has to adopt a wait and see policy. 28...♘d8 29 ♘h2 ♗g7 30 f4 f6 31 f5 Now White has a space advantage, not only in the centre but also on the kingside. 31...g5 32 ♗c2 ♗f7 33 ♘g3 ♘b7 34 ♗d1 h6 35 ♗h5 It’s fascinating to see how Karpov exploits the edges of the board to achieve his strategic goals. Now Black cannot prevent the exchange of the light-squared bishops. 35...♕e8 36 ♕d1 ♘d8 37 ♖a3 ♔f8 38 ♖1a2 ♔g8 39 ♘g4 ♔f8 40 ♘e3 ♔g8 41 ♗xf7+ ♘xf7 42 ♕h5 ♘d8 43 ♕g6 Only now does White agree to an exchange – when it improves his pawn position. All in accordance with the methods of Capablanca and Fischer! 43...♔f8 44 ♘h5 Black resigned. A strategical masterpiece by Karpov showing how to manoeuvre with the bishops at the very highest level. Learn the game from the beginning by heart! 83 Ernst – Engqvist Stockholm 2017 Black to move What’s so beautiful about chess is that you can steal ideas from its gods without doing anything morally wrong – and that is exactly what I did with my next move 20...♗xh2!. First I calculated 20...♖xh2 21 ♖xh2 ♗xh2 (21...♖f1+ 22 ♔c2 ♖xa1 23 ♖h7) 22 ♔e2 ♗f4 23 ♖h1 but in both variations I felt White had good chances to draw despite the fact that he was a pawn down. Then I analysed the text move which I immediately associated with the famous bishop move to a7 played in the abovementioned Karpov – Unzicker game. The point of the move in Karpov’s game was to double rooks on the a-file behind the bishop. Here I was happy to apply the same idea since I wanted to keep the game alive by retaining rooks on the board and playing for an attack. 21 ♔e2 ♖fh5 22 g4 ♖h3 23 ♖af1 f6 24 ♖f3 ♖3h4 25 g5 f5 26 ♔d3 ♗d6 By now I had little time on the clock so I agreed to the exchange of one pair of rooks. A good alternative was 26...♔d7. 27 ♖xh4 ♖xh4 28 b3 f4 29 ♗d2 29 ♗f2 ♖h5 and the g-pawn is lost. 29...♔d7 30 c4 c6 31 cxd5 31 c5 ♗c7 32 ♗e1 ♖h8 33 ♗d2 ♖f8 and Black wins. 31...cxd5 32 b4 ♔e6 33 a4 ♔f5 34 b5 ♔xg5 White resigned. 84 Halkias – Sokolov Olympiad, Bled 2002 Black to move 23...a5! White has played in accordance with the Karpov model and tried to take control of the c-file by first blocking it with the bishop deep down the file and then double his heavy pieces behind it. However, the problem is that in this position the concept doesn’t work since Black can just open another file. 24 ♘d3 24 bxa5 ♗xa3 or even worse 24 ♗xa5? b6 only make Black happy. 24...axb4 25 axb4 ♕a4 The a-file has been opened and Black exploits it while rendering the bishop manoeuvre to c7 completely useless. 26 h3 ♕b5 27 ♔f2 ♖a8 28 ♕c3 ♖c8 29 ♕a3 ♘f6 30 ♔e2? According to Flear the position would have been equal after 30 ♖c3. 30...♗xb4! Here the drawback of White’s play is most clearly seen. 31 ♕xb4 ♕xb4 32 ♘xb4 ♘e8 33 ♖a1 ♖xc7 White had to fight hard for a draw but eventually lost. A manoeuvre such as bishop to c7 is only really effective when there is just one open file and the position is locked as in Karpov – Unzicker. 85 Botvinnik – Tal World Championship match, Moscow 1961 White to move 35 ♗c4!? White’s plan is simply to play ♖h1, ♗d3, f4 and e5 with an unstoppable attack. 35 ♗xb7!? wins a pawn but slows down the pace of White’s positional plan. According to Philidor a pawn is worth three tempi in the opening. Here White loses two tempi, one to take the pawn and one to return to base. Philidor’s tempo rule hasn’t the same validity in the middlegame since everything is much more complicated, due to the fact that all the pieces have been mobilised. Yet the question remains whether it is worth two tempi to grab the pawn and apparently Botvinnik didn’t think so. A flexible move, and probably the most precise, was 35 ♖h1!, keeping Black in the dark somewhat about the light-squared bishop’s intentions. Botvinnik won quickly after 35...c5 36 b5 ♗f6? (36...a6) 37 f4 d3 38 ♖xd3 ♖xd3 39 ♗xd3 ♗d4 40 e5 g6 41 ♖h1 ♔g7 42 ♕e4 b6 43 ♗c4 Here the game was adjourned but Tal resigned without resuming the game. There is nothing to be done about the march of the e- and f-pawns. 86 Bogoljubow – Alekhine World Championship match 1929 White to move Superficially this position looks more equal than it really is since things are not quite what they seem. The reason is the bishops of opposite colours. This means that the player who takes the initiative has an extra piece to play with, since the opposing bishop cannot defend the same coloured squares of his own bishop. 19 ♗xb7? ♖ab8 leads to a position where Black has the initiative after for example 20 ♕a6 ♕xa6 21 ♗xa6 ♖xb2 with pressure on f2. The natural move 19 ♖ac1 only leads to further exchanges after 19...♖ac8. Correct is 19 ♖ad1! planning ♖d3-f3 with pressure on f7. 19...♖ac8 20 ♖d3! ♖c7 21 b3 If 21 ♖f3 then 21...♕b6. 21...b5 22 ♖f3 White is first to exert pressure against f7 since Black doesn’t have corresponding pressure on f2. 22...♕b6 23 a4 bxa4 24 bxa4 ♕b4 25 ♕a2 ♕b2 Alekhine makes a transition to an ending. 26 ♕xb2 ♗xb2 27 ♖b1 ♗d4 28 ♖b5 ♖a7 29 h4 h5?! Black should keep his pawns on the dark squares. 30 ♖b7! This exchange is in White’s favour since Black has to take care of two weaknesses on f7 and a5. 30...♖xb7 31 ♗xb7 ♖d8 32 ♗d5 ♖d7 Black’s rook has managed to move from the passive square f8 to a more active one but it remains passive compared with White’s rook, which has more manoeuvrability. 33 ♖b3 ♔f8 34 ♖b5 ♖a7 35 ♗b7 Black’s rook is temporarily trapped and this isn’t a nice position for Alekhine who lacks counterplay. It’s interesting that the computer evaluates the position as 0.00 which only goes to show that computers do not feel either positional or psychological pressure! 35...♔e7 36 g3 ♔d6 37 ♔g2 ♔c7 38 ♗d5 ♔d6 39 f4! f6? A better defence was 39...exf4 40 gxf4 g6 41 e5+ ♔e7 42 ♔f3 ♔f8. 40 ♗b3? Stronger was 40 f5! to isolate the h5-pawn by artificial means. With the move played Bogoljubow prepares ♗d1 to put pressure on the h5-pawn. Any ...g6 by Black can be answered by a timely f4-f5. 40...♖a6? Correct is 40...♗c3!. 41 ♗f7 ♖c6 42 ♖d5+! ♔e7 43 ♗xh5 ♖c5 44 fxe5 fxe5 45 ♗f3! ♔f6 After 45...♖xd5 46 exd5 White is practically two pawns up since the e5-pawn doesn’t mean anything. The possibility of creating two passed pawns is normally enough to win endings with bishops of opposite colours as we will see later in the endgame section. 46 ♖d6+ 46 ♖xc5? ♗xc5 is obviously a draw. 46...♔e7 47 ♖g6 ♔f8 48 ♗h5! ♗e3 49 ♔f3 ♖c3 50 ♔e2? 50 ♖e6 g5 51 ♔g4 gxh4 52 gxh4 is winning. 50...♗c5? Bogoljubow suggested 50...♗d4! with drawing chances. 51 ♖e6! White wins another pawn so the game is effectively over. It concluded: 51...g6 52 ♗xg6 ♖e3+ 53 ♔d1 ♖xg3 54 h5! ♔g7 Black must prevent h5-h6. 55 ♖xe5 ♗b4 56 ♔e2 ♖a3 57 ♗e8 ♖h3 58 ♗d7 ♖c3 59 ♖e6 ♗f8 60 ♖a6 ♗b4 61 h6+ ♔f7 62 ♗e6+ ♔f6 63 ♗f5+ ♔f7 64 ♖a7+ ♗e7 65 ♖xa5 ♖c2+ 66 ♔d3 ♖b2 67 h7 ♗f6 68 e5 ♗g7 69 ♖a7+ ♔f8 70 e6 ♗e5 71 e7+ Black resigned. A very instructive example which shows how to play with bishops of opposite colours. Even a small initiative means a lot between humans, especially if the defending player isn’t up to the task. 87 Englisch – Steinitz London 1883 Black to move Here Steinitz showed that if Black has the bishop pair in a symmetrical position with an open centre, there is a three-step-plan to exploit his advantage. Black’s main goal is to immobilise White’s minor pieces, especially the knight. Step 1 is to remove the outposts from the knight, in this case the d4-square. Step 2 is to drive the knight to a passive position with the help of the rook pawn. Step 3 is to open the game and exploit the passive knight. Steinitz played 17...b6!. This important move immobilises the bishop on the e3-a7 diagonal while the b3-knight cannot go to c5 or a5. 18 h3 ♗e6 Notice how both bishops exert pressure on the white queenside. 19 ♖fd1 19 ♘d4 would have been met by 19...♗c4 followed by ...c7-c5. 19...c5! This move concludes step 1 in the three-step-plan laid down by Steinitz. The white knight can no longer reach the outpost in the centre. 20 ♗g5 f6! It doesn’t matter that Black’s king’s bishop for the moment is temporarily passive. The main mission is to immobilise White’s minor pieces. 21 ♗f4 ♔f7 A typical advantage of the bishop pair is that it’s often easier to activate the king than it is for the defender in this type of position. If White’s king goes to f1 Black can check on c4 and if White prepares f3 and ♔f2 White must look out for a check on e2 by a rook defended by a bishop on c4. 22 f3 g5! White’s bishop is immobilised on the c1-h6 diagonal as well. 23 ♖xd8 ♖xd8 24 ♗e3 h6 25 ♖e1 f5! Black opens the long diagonal for his bishop and threatens the positionally desirable ...f5-f4. 26 f4 ♗f6 27 g3 a5! This is the initiation of the second phase in the three-step-plan: to immobilise White’s knight. Black’s threat is to play ...a4 followed by ...a3 when White’s queenside pawns will be destroyed. 28 ♘c1 a4 29 a3 29...♗c4! This is also a typical move in positions where the bishop is stronger than the knight. Black’s c4-bishop dominates the c1-knight. This ends step 2. Step 3 is to open the position and exploit Black’s more active position. 30 ♔f2 gxf4 A more typical plan was 30...♖d5 to prepare the decisive break ...b5-b4. Nevertheless Steinitz’s continuation is good enough since all kinds of open positions benefit Black who has a superior position where the bishop pair is a decisive factor. 31 ♗xf4 31...♗g5! One advantage of possessing the bishop pair is that it’s easy to decide when it’s time to make a profitable exchange. It’s not necessary to retain the bishop pair to the end, since sometimes it is important to take advantage of its dynamic function. If Black wanted to keep the two bishops he could always have played 30...♖d5 instead. 32 ♗xg5 hxg5 33 ♔e3 ♔f6 34 h4 34 ♖h1 ♔e5 followed by ...f5-f4 decides the game in Black’s favour since the rook enters d2 after the check with the pawn. 34...gxh4 35 gxh4 ♖e8+ 36 ♔f2 ♖xe1 37 ♔xe1 ♔e5 White’s distant passed pawn is effectively neutralised by Black’s active king. 38 ♘e2 ♗xe2 39 ♔xe2 ♔f4 40 c4 ♔g4 41 ♔e3 f4+ 42 ♔e4 f3 43 ♔e3 ♔g3 White resigned. This is a very important model game showing how to play and win with the bishop pair. There is another quality game between Rosenthal – Steinitz, Vienna 1873, which the reader should study as it illuminates the same three-step-plan. Don’t miss it! This ends the section handling manoeuvring with minor pieces. We now move on to major pieces, beginning with the rook. 88-98: The Rook Manoeuvre The rook is probably the most difficult piece to manoeuvre. In the beginning of the game and in the early middlegame it’s sometimes impossible to know where the rook is best placed from a long-term perspective. In the endgame there are many open files and ranks and it’s not always easy to calculate the correct continuation. 88 Kovačević – Seirawan Wijk aan Zee 1980 Black to move If players like Ivanchuk and Shirov seem to come from another tactical planet the same could be said about Seirawan regarding his positional play. It looks like White is controlling the h-file but in reality it’s Black who will take over that file after the highly original 14...♔d7!!. This creative move prepares not only ...♖h8 but also a king manoeuvre to c7 with the result that Black’s position becomes more harmonious. The game continued: 15 ♘bd2 ♖h8 16 ♖g1 ♔c7 17 ♖b1 ♖h3 18 b3 ♕h8 19 ♘f1 19 ♗f1 was stronger. 19...♘8d7 20 ♗f4 ♘e5 21 ♘xe5 ♗xe5 22 ♗xe5 ♕xe5 Black’s queen and rook dominate the position. 23 f3 ♗d7 24 ♕c2 ♕d4 25 ♖g2 ♖h1 26 ♖f2 26...♕h8 The trivial 26...♗h3 was better but remember Seirawan is from another positional planet! 27 f4 27 ♕d2 followed by ♕e3 was a better defence. 27...♕h4 28 ♖d1 f6 29 gxf6 exf6 30 e5 fxe5 31 fxe5 ♖f8 32 exd6+ ♔b7 33 ♗d3 ♖e8+ White resigned. Seirawan and Petrosian must be regarded as the most original players when it comes to handling their king. 89 Blackburne – Nimzowitsch Saint Petersburg 1914 Black to move 14...♖e8! The idea of this mysterious rook move is to make it harder for White to play d3-d4. 15 ♘h4 ♗f8 16 ♘f5 ♔h8 17 g4 ♕d7 18 ♘f2 18 g5 g6 19 ♘h4 f5!. 18...a5 19 a3 b5 20 ♖ad1 ♖ad8 21 ♖d2 b4 22 axb4 axb4 23 c4? White gives up the d4-square without any fight. Now the strongest continuation according to Nimzowitsch was 23...g6 24 ♘h4 ♘d4! 25 ♗xd4 exd4 followed by ...♗h6. 90 Bronstein – Petrosian Moscow 1967 White to move Bronstein played the mysterious rook move 16 ♖ac1. Petrosian replied 16...♘d5. The idea behind White’s 16th move is made very clear in the variation 16...g4 17 ♘d4, since the c2bishop is defended after a black trade on d4. 17 b4 This pawn advance was also made possible by the prophylactic rook move. 17...♘f4 18 ♗xf4 gxf4 19 ♘f5 ♗f8 Or 19...exf5 20 b5. 20 b5 ♘e7 21 ♘xe7 ♗xe7 22 ♘e5 ♗d6 and now the best was 23 ♕h5 ♗xe5 24 ♖xe5 ♕xc3 25 ♖c5 with a clear advantage. 91 Karpov – Kasparov World Championship match, Moscow 1985 Black to move It’s not easy to decide how to develop the rooks in this position. Black wants to expand on the queenside with ...b7-b5 very soon and then b8 would be an attractive square. But if …b7-b5 is met by axb5 you want a rook on a8. The c-file is semi-open and so it’s natural to place a rook on c8. The d8-square also looks useful since most of the time it’s good to have a rook behind a central pawn, especially if it advances. Black can prepare either ...d6-d5 or ...e6-e5 and also be ready in case White plays e4-e5. The e8-square is also interesting since Black can counter White’s f4-f5 advance and prepare ...e6-e5 when the pawn is placed on f4. It’s also good to have a rook on f8 when White goes for g4-g5, since the break ...f7-f6 might become important. Kasparov played 11...♖e8!. In this position it makes sense to place the rook on e8 in accordance with the general rule that rooks are effective behind advancing pawns. 12 ♗f3 ♖b8 Black moves the rook away from the long white diagonal and prepares a future ...b7-b5. 13 ♕d2 ♗d7 14 ♘b3 b6 15 g4 ♗c8 The bishop evacuates the d7-square for the knight and prepares a later ...♗b7. 16 g5 ♘d7 17 ♕f2 ♗f8 18 ♗g2 ♗b7 19 ♖ad1 g6 20 ♗c1 ♖bc8 The rook no longer has any function on b8 but according to Kasparov it’s a pointless waiting move. He wants to improve with 20...♘c5 hoping for 21 ♘xc5 bxc5 when the rook is well placed on b8. 21 ♖d3 ♘b4 22 ♖h3 ♗g7 23 ♗e3 ♖e7! 24 ♔g1 92 Black to move 24...♖ce8!! Black understands that White wants to break through with f4-f5 and so prepares to make this less appetising by doubling rooks on the e-file. It’s an unbelievable rook manoeuvre never to be forgotten. It might seem strange to double rooks on a closed file but this is how to use rooks at a higher level. Incredibly skilful play by Kasparov even if it’s a slight exaggeration to claim that he became World Champion because of his ability to play with the rooks! Interestingly, though, Kasparov attached the punctuation ?! to his move and sought to improve with the immediate 24…f5. 25 ♖d1 25 ♕h4 ♘f8 is nothing for White. 25...f5! 26 gxf6 26 exf5? exf5 obviously activates both Black’s rooks. 26...♘xf6!? Kasparov believes that this unexpected pawn sacrifice won the match rather than his skilful play with the rooks. Objectively speaking, 26...♗xf6 was safer, but perhaps it would not have won the game. This is what calculated risk is all about! 27 ♖g3 ♖f7! Black’s rook no longer has any function on the e-file and so moves to the f-file. Observe Kasparov’s flexibility with the rooks and how he always adapts to new situations arising on the board. 28 ♗xb6 ♕b8 29 ♗e3 ♘h5 30 ♖g4 ♘f6 31 ♖h4? g5! 32 fxg5 ♘g4 33 ♕d2 ♘xe3 34 ♕xe3 ♘xc2 35 ♕b6 ♗a8 36 ♖xd6? 38 ♕xb8 ♖xb8 was better with mutual chances. 36...♖b7! Black’s king’s rook triumphs and decides the game. This encounter secured the World Championship title for Kasparov. 93 Gelfand – Illescas Madrid 1996 Black to move White can play for c4-c5 to open files on the queenside or d4-d5 to immobilise the black bishop and activate the knight on d4. It’s also possible to manoeuvre the knight to d2 and trade it for Black’s centralised knight. 12...♖f6! A strong attacking plan which prepares the rook manoeuvre to h6. Since Black has an attack in mind against White’s king one can, in accordance with Suetin’s principles, say that the middlegame starts exactly here. The normal plan is 12...♕e7 followed by ...e5 with play in the centre. 12...g5? followed by ...g4 is far too risky since it weakens the kingside too much. 13 d5 ♖g6! 13...e5 is met by 14 ♘h4!. 14 dxe6 14 ♘e1 or 14 ♘d4 were the alternatives. 14...♘f8! 15 c5 15 ♘e1!? 15...♘xe6 16 cxd6 cxd6 17 ♖ad1 ♔h8 18 ♘e1?? A natural but fatal move! 18 ♔h1 should have been played. 18...♘6g5! White cannot stop the knight from coming to h3 with decisive effect. 19 ♔h1 ♘h3! ...and White resigned due to 20 gxh3 ♘xf2 mate. 94 Tarrasch – Lasker World Championship match, Düsseldorf/Münich 1908 Black to move Black’s position looks rather inconspicuous and is a good example of how one can turn “nothing” into “something” with the help of piece activity. Lasker played the impressive move 15...♖e5!. This is the most active as well as the best move, at least from a psychological point of view. It represents the second phase of development where some of the pieces try to penetrate deeper into the position. This is normally something for the knight but here this mission is accomplished by the rook which seeks play on the rank rather than the file. There were several candidate moves such as 15...♘c5 (the computer’s preference) 15...♘f8, 15...♘b6 and 15...a6 with the cheapo 16 ♘xc7?? ♖c8. However, all of these continuations must be regarded as slightly more passive than Lasker’s choice. Lasker understood very well the value of using the rook along the rank, especially when playing against mere mortals! 16 ♘d4 The point of Black’s original and creative rook move wasn’t just to protect the c7-pawn (16 ♘xc7?? ♖c5) but also to be able to play: 16...♖c5!. A very thoughtful concept by Lasker. White already has two rooks in play and Black tries to compensate for this by playing actively with his only developed rook and by using the e5-square as a pivot (manoeuvring square). Black’s e-file is actually longer than White’s and this fact is exploited to the maximum by Lasker. 17 ♕b3 ♘b6 The knight helps to create an outpost for the rook on c4. 18 f4 Tarrasch plays according to classical principles and tries to cut off the rook from the centre, the e5-square. However, Lasker, who played beyond the classical principles developed by Steinitz and Tarrasch, has other plans for this rook. 18 ♖e3 could have been met by 18...♕e8 19 ♘f5 ♔h7. 18...♕f6 19 ♕f3 ♖e8 Black’s rook number two is in the first phase of its development but here it shows itself to be useful by exerting pressure on the centre. 20 c3 a5 21 b3 a4! 22 b4 ♖c4 With imaginative play Lasker has created a perfect outpost for his rook where it can apply pressure on the c3-pawn and eventually also on the e4-pawn after an appropriate pawn push ...c7-c5. 23 g3 23...♖d8 Black prepares ...c7-c5 but releases the pressure on the e4-pawn. An immediate 23...c5? is premature because of 24 ♘b5 with a clear advantage to White. 23...♕e7!? or 23...♕g6!? were playable moves with a focus on e4. 23...a3!? was also interesting, planning to exert pressure on c3 after ...♘a4. 24 ♖e3? 24 ♕d3 or 24 ♖d2 was better with a small advantage to White. 24...c5! Now Lasker takes over the game. 25 ♘b5?! 25 ♘c2 or 25 bxc5 with an equal game was better. 25...cxb4 26 ♖xd6? 25 ♖ed3 is better. 26...♖xd6 27 e5 27...♖xf4! It’s significant that Lasker decides the game with the same rook he has managed so creatively. Presumably it was this move that Tarrasch overlooked with his 24th, 25th or 26th move. 28 gxf4 ♕g6+ ...and Lasker won the game after the further moves: 29 ♔h1 ♕b1+ 30 ♔g2 ♖d2+ 31 ♖e2 ♕xa2 32 ♖xd2 ♕xd2+ 33 ♔g3 a3 34 e6 ♕e1+ 35 ♔g4 ♕xe6+ 36 f5 ♕c4+ 37 ♘d4 a2 38 ♕d1 ♘d5 39 ♕a4 ♘xc3 40 ♕a8+ ♔h7 41 ♔h5 a1♕ White resigned. 95 Gligorić – Stein Lvov 1962 Black to move Stein was a true master when it came to reaching positions with strong tensions. Superficially it might seem that White’s passed pawns on c4 and d5 are more dangerous than Black’s on d4 and e5, since the bishop on e4 helps to blockade them and can support its own with the c4-c5 push at the right moment. However, it’s not that simple. The natural continuation 18...♖ac8 19 b3 b5 has no point after 20 ♖ac1 bxc4 21 bxc4 ♗xe4 22 ♕xe4 ♖xf1+ 23 ♖xf1! ♕d6 24 ♖c1 and White’s pieces are better than Black’s. Stein played 18...♖f6! planning a doubling of the rooks and increasing the tension on the f-file. 19 ♖f3! ♖af8 This is where the black queen’s rook really belongs, behind its colleague on the f-file, rather than the c-file. 20 ♖af1 h5 21 c5 21 ♗g5 leads to an equal game after 21...♗xe4 22 ♕xe4 ♖xf3 23 ♖xf3 ♖f5!. Note how Black retains the tension with the rook. However, if White releases the tension this would leave Black with positional gains. 21...♕b5! A beautiful positional move, threatening to win with ...d4-d3. 22 ♗xf5 22 a4 ♗xe4 23 ♕xe4 ♖xf3 24 ♖xf3 and the position will be equal no matter which pawn Black chooses to take 22...gxf5 23 ♗g5 23 ♖xf5?? d3. 23...♖6f7 and Stein won this highly dramatic game after 57 moves. 96 Bologan – Ruzele Kramatorsk 1989 White to move As a rule, positions where White has an isolated d4-pawn versus a c6-pawn are better for White than playing with an isolani against an e6-pawn. The main reason is that White’s bishop, stationed on the classical diagonal, has greater scope all the way to the vulnerable f7-pawn. If Black places his knight on d5 the bishop can switch to the b1-h7 diagonal. White can easily exploit the open e-file and use the e5-square as an aggressive outpost. Despite the fact that Black has the open c8-h3 diagonal at his disposal the bishop is often unable to find a suitable square. Bologan played the strong 21 ♖e5! which provokes an exchange as the rook is in a strong position where it threatens a doubling of rooks as well as taking twice on d5. 21...♖xe5 21...f6 22 ♗xd5+ cxd5 23.♖xd5 ♗c6 24 ♖d6 ♗xf3 25 gxf3 and White has a sound extra pawn. 22 dxe5 White has used Philidor’s method to improve his pawn structure, i.e. the outpost on e5 was exploited to make possible an exchange which at the same time improved his pawn structure. White’s isolated pawn is certainly artificially isolated on e5, since it is far advanced compared with his other pawns, but it’s easy to protect on e5 and it has an aggressive position which offers dangerous outposts on f6 and possibly d6. White’s next goal is to manoeuvre the knight to f6. 22...♗e6 23 ♖e1 The e5-pawn is overprotected to make the mini-manoeuvre ♘g5-e4 possible. 23...h6 24 ♘d2 The preparation for the decisive manoeuvre to f6 (d6). 24...♕f8 25 ♘e4 ♕b4 26 ♖e3 ♔g7 26...♘xe3?? 27 ♘f6+ and White wins the queen. 27 ♗xd5 cxd5 28 ♕f6+ ♔g8 29 ♘d6 ♕b6 30 ♖g3 White is threatening a sacrifice on g6. 30...♔h7 31 ♘xf7 ♖g8 32 ♖xg6! Black resigned. Note that the principal reason why all White’s pieces ended up so aggressively placed originated with his 21st move. 97 Reinhardt – Lombardy Mar del Plata 1958 Black to move If we ignore the d4-square White has in principle only one other possible weakness and that is the a2-pawn. Black’s general plan is therefore to play ...♖a6 followed by ...♖a4 and ...♕a5 to force b2b3. After that Black can prepare the break ...a5-a4. 16...♖b6! 17 ♖c1 ♖a6 Many chessplayers have discussed this remarkable idea of Lombardy but it actually originates from the chess master Sidney Bernstein. 18 b3? It’s better to go back with the rook. One plausible sequence according to Lombardy is 18 ♖a1 ♘d4 19 ♗xd4 ♗xd4 20 ♘e3 ♗g7 21 f5 ♖a4 22 g5 f6 and the game is in a state of dynamic balance. Black will at the right moment play his queen to a5 to force b2-b3. The whole point of Black’s play is to make the a2-pawn a problem for White. 18...♘d4 19 ♗xd4 ♗xd4 20 ♖c2 ♖a3! 21 ♘e3 a5 22 g5 a4 23 f5 axb3 24 axb3 ♖xb3 Black has broken through and won a pawn, whereas White’s attack isn’t at all dangerous. The game concluded: 25 ♘g4 ♖a3 26 h4 b3 27 ♖b2 ♕a5 28 ♘h6+ ♔g7 29 ♖bb1 b2 30 fxg6 hxg6 31 ♕f3 ♗e5+ 32 ♔h1 ♖a1 33 ♖bd1 ♕b4 34 ♖b1 ♕d2! Black is approaching White’s king via the queenside. 35 ♕d1 ♖xb1 36 ♕xb1 ♗h3 37 ♖g1 ♕f4 White resigned. It should be mentioned that this plan impressed Fischer so much that he used it in one of his own competitive games. 98 Carlsen – Ivanchuk Foros 2008 White to move 28 ♖a1! Carlsen believes the rook belongs on a2 where it can prepare a4a5 and manoeuvre to c2 in some variations. This last mentioned idea to get out of the firing line of the bishop on the h6-c1 diagonal, stems from a game Petrosian – Gufeld, played in the Soviet Championship 1960. In that game Petrosian kept clear of the opponent’s dark-squared bishop along the h6-c1 diagonal by playing the moves a2-a4, b2-b3, ♖a1-a2-c2, ♖f1-b1-b2. Notice that by playing on the second rank with the rooks, the last rank is “removed” and so the board is reduced to seven ranks instead of eight. It’s because of such model examples that Carlsen, with his extraordinary memory, can make use of the most important classical ideas from the past. The game continued 28...♕h6 29 ♖a2! ♘f6?! Ivanchuk didn’t realise the danger on the c-file because it’s important to block it with a timely ...♘c5. 30 ♔g2 ♘h5?! 30...♘d7 is better. 31 ♘d5! ♘xd5 32 ♗xd5 ♗f4 33 ♕f2 fxe4 34 ♗xe4 ♕g5 35 ♖c2! 35...d5? 35...♖f7 was the best but White has the incredible beautiful bishop manoeuvre 36 ♗b7!! with the idea of ♗c8. Note that White exploits the fact that Black has reduced the board to seven ranks. 36 ♗xd5 ♗xg3 37 hxg3 ♘f4+ 38 ♔f1 ♘xd5 39 ♖ce2! ♕f6 39...♖f5 is answered by 40 g4 or 39...♖e8 by 40 f4. 40 ♖xe5 ♕xf3 41 ♕xf3 ♖xf3+ 42 ♔e2 ♖f5 43 ♖xf5 gxf5 44 ♔d3 c5 45 ♖e5 ♘b4+ 46 ♔d2 Black resigned. Let’s now move on to queen manoeuvres. 99-102 The Queen Manoeuvre The queen is the strongest tactical piece on the board and therefore it’s a good idea to centralise it whenever possible. When there are no targets around it must be content with controlling a file and acting like a rook, or controlling a diagonal and assisting or replacing a bishop. However, when there are targets like unprotected pieces, weak pawns or an exposed king, it shows its teeth and then you think: this isn’t a queen, it’s an octopus. 99 Petrosian – Barcza Budapest 1955 White to move 9 ♕d1 Clarke wrote in his comments: “Such a move shows a far greater understanding than any brilliancy. The queen has already fulfilled its purpose in going to f3 and has nothing else to do there.” The queen goes back to d1 to help defend the weakened queenside. The natural developing move 9 ♘ge2 might have been answered with 9...♘d7 when White has to figure out how to meet ...♕b6. 9...♕b6 10 ♕d2 ♘a6 11 ♘f3 0-0-0 12 a3 The preparation of a minority attack on the queenside. Of course this is far more dangerous when Black’s king is on the same side as the white minority attack. It’s impossible for Black to do likewise on the kingside and the g-file is utterly useless. 12...♘c7 13 b4 ♘e8 14 ♗e2 ♘d6 Black’s knight is placed on its best square but it will not compensate for the other defects in Black’s position. 15 ♕a2 A nice queen move that takes the sting out of a possible ...♘c4. 15...♘e4 16 ♖c1 ♘xc3 17 ♖xc3 ♗d6 18 0-0 ♔d7? The king should have headed for the queenside with 18...♔b8. 19 ♘h4 White endeavours to control the f5-square. 19...♔e7 20 f4 ♗e4 21 ♗d3 ♗xd3 22 ♖xd3 ♖d7 23 ♘f5+ White is strategically dominant and can win as he pleases. Black resigned after 60 moves. 100 Winter – Alekhine Nottingham 1936 Black to move 23...♖8e7! This move prepares to triple on the e-file and is called “Alekhine’s Gun.” It means that the queen will be placed to the rear of the two rooks. The effect will be devastating and this is the strongest pressure that can be created along a file. Note that diagonal pressure can only be created by a battery of queen and bishop. 24 ♔d1 White prepares ♘g1. An immediate 24 ♘g1 is met by the finesse 24...♗xf4! 25 ♖xe6 and now the intermediate move 25...♗xd2+. 24...♕e8 25 ♕f3 25 ♘g1 ♘g3 26 ♖xe6 ♖xe6 27 ♖e1 results in an exchange of rooks but it doesn’t help White since Black’s position is too strong after 27...♘e4 28 ♔c2 ♘e7 followed by ...♘f5 when all Black’s pieces are active. 25...♘a5! 26 b3 26 ♕xd5 ♖xe2 27 ♕xa5 (Or 27 ♖xe2 ♖xe2 28 ♕xa5 ♘g3 29 ♖f3 ♕e4 with a mating attack.) 27...♘g3 with a winning position. 26...♘c4! The b-pawn is no hindrance for Black’s knight move to c4. 27 ♗c1 27 bxc4 ♕a4+ 28 ♔c1 ♗a3+ 29 ♔b1 ♖b6+ 30 ♔a1 ♕c2 and White is mated in two moves. 27...♘ce3+ 27...♕b5 first was even stronger. 28 ♗xe3 ♖xe3 Also 28...♘xe3+ 29 ♘xe3 ♖xe3 30 ♕f2 ♖d3+ 31 ♔c2 ♖ee3 won due to the devastating ...♕e4. It’s not often one sees such a centralisation of the major pieces! Note the doubling on the sixth rank from Black’s point of view. 29 ♕f2? 29 ♘xe3 ♘xe3+ 30 ♔d2 ♘xf1+ 31 ♖xf1 ♖e4 would have prolonged the game. 29...♕b5 30 ♘c1 ♖xc3 and Black won. 101 Capablanca – Berasain Montevideo 1911 White to move 27 ♕d3! The reasoning behind this move is that Capablanca had discovered that it’s more effective to have the rook on d5 instead of the queen – and have the queen behind the rook. 27...♕e6 28 ♕e2 ♖c6 29 ♖d5 Note that the queen and rook are manoeuvring back and forth in this game and how they are exploiting d5 as a manoeuvring square, a pivot, to use a term of Nimzowitsch who was the first to document the manoeuvre in his famous classics. 29...♖e8 30 ♕d3 The regrouping is now over. Capablanca has set up “Alekhine’s gun” reversing the positions of his queen and rook. Normally it’s best to have the queen behind the two rooks and second-best is to have it behind the first rook. In this position Black must reckon with the threat on d6. 30...♕g6 31 ♘e3!? The problem with Black’s position is that it suffers from immobility. In fact White wins the game by exploiting his own superior mobility to carry out purposeful manoeuvres. White doesn’t win a pawn after 31 ♘xd6? cxd6 32 ♖xd6 ♖xd6 33 ♕xd6 because of 33...♕xe4 – for example 34 ♕xc5 ♕c2 35 ♖a1 ♕b2. A sound alternative was to establish a pawn wave on the kingside with 31 g3 followed by h2-h4. 31...♖b8 32 ♕c4 The c4-square is also an important pivot for White’s knight and queen. 32...♖cb6 33 ♘f5 ♔h7 Black places his king out of the range of the knight to avoid a devastating check on e7 after a conceivable sacrifice on d6. 34 ♖5d3! ♖g8 35 ♖g3 ♕f6 36 ♘e3! The knight is on its way to d5. Soon all of White’s pieces will have visited the d5-square! It’s typical that pieces change places when there are several manoeuvring squares available. 36...♕e6 37 ♕xe6 White sacrifices the d5-square but in exchange he will have the c4-square and the clearly better pawn structure. The black bishop is like a Russian pawn! 37...fxe6 38 ♘c4 ♖a6 39 ♖d2 ♖b8 40 ♖gd3 ♖d8?! 40...g5 followed by ...♔g6 was a better defence even though it will not hold in the long run. 41 ♘xe5 ♖b8 42 ♘c4 ♗f4 43 ♖b2 43 ♖e2? ♖xb3 44 g3 ♖b1+ 45 ♔g2 ♖a4! and White has lost control of the fine manoeuvring square on c4 and also has some weak pawns on the queenside. 43...♗c1 44 ♖c2 ♗g5 45 g3 ♗d8 46 ♖cd2 Black resigned. A beautiful and almost perfectly played game by Capablanca. It shows not only how to manoeuvre with the queen but also how to manoeuvre with the rooks and the knight and how these pieces make good use of the important manoeuvring squares on d5 and c4. 102 Karpov – Kasparov World Championship match, Moscow 1984 Black to move Kasparov writes the following after the strong queen manoeuvre 16...♕e8! “Black has set up an original battery on the e-file and in the best traditions of the Tarrasch Defence he does not intend to defend the d5-pawn with all his might, but wants to initiate play on the c- and e-files, exploiting the activity of his light-squared bishop.” Note that all Black’s pieces are active with the exception of the f6-knight which has a defensive function. The bishop on g4 exerts pressure on the e2-pawn and this is against Nimzowitsch’s advice that you should place the bishop on e6. In modern chess it’s activity that counts and this is the sole idea of the Tarrasch Defence or whenever you are saddled with an isolani in the middle of the board. Kasparov is playing according to the spirit of the system and therefore applies pressure on the central files as well as controlling the important outposts on c4 and e4. 17 h3 17 ♗xd5?! is strongly answered by 17...b5, for example 18 a3 (18 b4 ♖xe3!) 18...♗h3 19 ♖fe1 (19 ♗g2 ♗xg2 20 ♔xg2 ♖xe3 21 fxe3 ♘c4) 19...♖xe3 20 fxe3 ♕xe3+ 21 ♔h1 ♕f2 22 ♕d2 ♘g4 23 e3 ♘c4 24 ♕e2 ♘cxe3. Note the activity in all Black’s pieces. This is how to play with an isolani! After 25 ♖d2 ♘xd5 26 ♕xf2 ♘xf2+ 27 ♖xf2 ♘xc3 28 bxc3 ♖xc3 it’s White who is happy with a draw. If White plays 17 ♗d4 the pressure is relieved on the d5-pawn and Black can activate his knight with 17...♘e4!. 17...♗h5 17...♗e6? blocks the e-file and upsets the harmony and spirit of Black’s position. Even worse is perhaps the fact that White can play 18 ♗xa7 winning a pawn for virtually nothing. An exchange sacrifice inspired by Petrosian or Andersson isn’t really necessary here and should therefore principally be avoided. 17...♖xe3 18 fxe3 ♕xe3+ 19 ♔h2 ♗e6 is possible but after 20 ♕c1 Black cannot avoid the exchange of queens. Great exchange sacrificers, like those players just mentioned, made such moves only when it was the best continuation. 18 ♗xd5?! This isn’t the best move but perhaps it can be regarded as pragmatic due to the fact that Karpov’s style is rather materialistic. The best move was 18 ♕f5 which might lead to equalisation after 18...♖xc3! 19 bxc3 ♗xe2. 18...♗g6 18...b5! might have been an even better way to activate Black’s position. A step in the wrong direction was a natural continuation like 18...♘xd5?! 19 ♖xd5 ♗g6 20 ♕d1! ♘c4 21 ♗c1, since it only helps to introduce some harmony to White’s game. In this position White’s queen and bishop are perfectly placed on their initial squares! 19 ♕c1 19 ♕xg6 ♘xd5 is obviously in Black’s interest, not only because he threatens the queen but also e3 and perhaps also c3. 19...♘xd5 19...♗e4 was playable to maintain the balance but would have led to a different kind of game, less dynamic. 20 ♕d2 is probably best, targeting the unprotected a5-knight. 20 ♖xd5 ♘c4 The quiet move 20...b6 was also good, as it frustrates the white bishop’s prospects on the g1-a7 diagonal. 21 ♗d4 The greedy 21 ♗xa7? b6 22 b3 is answered by the strong 22...♘e3! 23 fxe3 ♖xa7 after which White’s pawn structure and knight become unstable. 21...♖ec7! Black fully exploits the potential of all his men – major and minor or pieces alike. Nevertheless Karpov still managed to win the game in 42 moves after Kasparov made some mistakes in the time scramble. We will now concentrate on positions where several pieces are involved in intricate manoeuvring. Not only by using manoeuvring squares but also by making use of the whole board as well as applying the principle of two weaknesses. 103-108 Manoeuvring with Several Pieces 103 Steinitz – Showalter Vienna 1898 White to move This position is both a forerunner and a basic example of how White can play on both wings simultaneously. White has a large advantage in terrain due to the strongly defended pawns on the fifth rank as well as the super knight on d4, hitting the e6-weakness. The best way to exploit the huge space advantage is by opening a file. The important question though is not what to do (c3-c4) but when to do it. Steinitz’s fundamental plan is to commence manoeuvring on the kingside in order to lure the pieces to that sector of the board and then open the queenside. This is in accordance with the concept of two weaknesses. 19 ♕f2! White prepares ♖e3-h3 and for the moment keeps 19 c4 in reserve, when it will be even more effective. It’s too early to play it now due to 19...♖c8, for example 20 ♖ac1 ♖c5. 19...♕d8 20 ♖e3 g6? Of course this unprovoked weakness makes White’s task easier. 21 ♖h3 ♖f7 22 ♔h1 22 ♘f3! was more precise. 22...♔g7 23 ♘f3 h6 24 ♖g1 White is now threatening the decisive break g2-g4. 24...h5 25 ♕g3 ♕h8 26 ♘g5 ♖ff8 27 ♕h4 ♘g8 27...♖a7 is answered by 28 c4!. 28 c4! The moment has come for this long awaited break but ironically there was an even better break, the “impossible break” – 28 g4!!, which would have won more quickly after 28...fxg4 29 ♖xg4. 28...dxc4 29 ♗xc4 ♖fe8 30 ♖d3 ♖a7 31 ♖d6 ♖b7 32 ♖gd1 ♗c8 33 ♘xe6+ ♗xe6 34 ♗xe6 ♕h7 35 ♖d7+ ♖e7 36 ♖xe7+ ♘xe7 37 ♕f6+ ♔h6 38 ♖d8 ♖c7 39 h3 Black resigned. 104 Rubinstein – Lasker Moscow 1925 Black to move Just like Capablanca – Berasain, position 101, Black will use the d5-square as a pivot for manoeuvring. Lasker played 27...♘b6 intending to use the d5-square for his other pieces and preparing ...♗d5, ...♕c6 followed ...a5-a4 to create another pivot on c4. 28 ♔h2 ♗d5 29 ♔g2 ♕c6 30 ♘d2 a5! 31 ♕c3 ♗xf3+! 31...♕xc3 32 ♗xc3 ♗xf3+ 33 ♔xf3!. 32 ♘xf3 32 ♕xf3 is met by 32...♘d5!. This is stronger than the immediate 32...♕c2 because of 33 ♕c3 when White cannot prevent the invasion on c2, for example 33 ♕e4 is met by 33...f5. 32...♕xc3 33 ♗xc3 a4 34 bxa4 34 b4 ♘c4 or 34 ♗a5 axb3 35 ♗xb6 and now the beautiful tactic 35...♗g5!! ensures that the b-pawn promotes. 34...bxa4 35 ♔f1 After 35 ♗b4 ♗xb4 36 axb4 a3 37 ♘d2 a2 38 ♘b3 ♔f8 Black wins the knight ending after ...♘d5xb4 etc. 35...♗xa3 36 ♔e2 ♔f8 37 ♔d3 37...♘d5 Note how Black uses the pivot on d5 to decide the game. 38 ♗e1 ♗d6 39 ♔c4 ♔e7 40 ♘e5 ♗xe5 41 dxe5 ♔d7 42 ♗d2 h5 43 ♗c1 43 ♔b5 a3 44 ♗c1 ♘c3+! 45 ♔c4 a2 46 ♗b2 ♘e4 47 ♔b3 ♘xf2 and Black wins another pawn on the kingside. 43...♔c6 44 ♗a3 ♘b6+ 45 ♔d4 ♔b5 46 ♗f8 ♘c4 47 ♔c3 g6 All black pieces are placed on white squares and that’s typical for a light-square strategy. 48 f4 ♘e3 49 ♔d3 ♘d5 50 ♗a3 h4 51 gxh4 ♘xf4+ 52 ♔e4 ♘h5 52...♘xh3?? 53 ♔f3 and Black’s knight will be continually under threat after ♔g3 and ♔f2. 53 ♔f3 Rubinstein’s last chance is to penetrate the black position and catch the f-pawn via g4-g5-h6-h7-g8. 53...♔c4 54 ♗b2 ♔b3 55 ♗a1 a3 56 ♔g4 ♔c2 57 ♔g5 ♔d3! White resigned. Black’s king can be placed on f5 after White has captured the f7pawn and then the knight manoeuvre to b3 or c2 wins. Note that the game will be decided on a black square (a1!) and that is also typical of a white-square strategy as Bronstein so insightfully once remarked. A very nice game played by Lasker who showed how to conduct a white-square strategy when the opponent has an isolani. It’s not necessarily the isolated pawn that is weak but rather the white squares around it. This concept has been very well elaborated by Nimzowitsch in his book My System. This encounter must be regarded as one of the most important model games to show how to manoeuvre when the opponent has a complex of weaknesses of the same colour. Study it thoroughly and you will be well rewarded in your own games. 105 Bogoljubow – Réti Berlin 1919 White to move The move 11 a4! looks like preparation for an attack on the queenside but the move primarily prepares a knight manoeuvre to c4. Later this knight will manoeuvre to g2 via e3 after White has played the g-pawn to g4. You might say that White’s pawn move on the left flank prepares an attack on the right! 11...a5 Obviously 11...♘f4 would have disrupted Bogoljubow’s grandiose plan but on the other hand White would have the possibility of gaining space on the queenside after 12 ♗xf4 exf4 13 a5. 12 ♘c4 b5? Black understandably doesn’t like the white knight pressurising d6 but nevertheless it would be better to prepare ...♘b6 by playing 12...♖e8. 13 axb5 cxb5 14 ♘e3 14 ♘a3! won a clear pawn on b5 but the win of a pawn doesn’t seem to bother Bogoljubow who is focused on the kingside! 14...♘b6 15 g4 This is the profound point of Bogoljubow’s 11th move. However, do note that 15 ♗xb5, objectively speaking, was a good alternative. 15...♘f4 16 ♗xf4 exf4 17 ♘g2 White’s original knight manoeuvre from c4 to g2 has been successful and White has the slightly better position. Most important here is not the fact that White could play stronger but rather that one should remember to watch out for moves on one side (11 a4) which can help to make progress on the other side (♘c4-e3, g4 and ♘g2). This is the ideal way to play over the whole board and Bogoljubow repeatedly showed that he was a master of this difficult kind of play. His method of starting an attack on one wing as preparation for an attack on the other was a new strategic idea and it’s only right to mention that Bogoljubow was the first to cultivate play over the whole board. In fact, as we have just seen, this was even more important for him than gain of material. I recommend a deep study of his best games so you can master the complicated art of playing over the whole board. Apart from Bogoljubow, other great attackers like Alekhine and Kasparov have also embraced this method of combat. 106 Jianjun Xie – Engqvist Shenzhen 2012 White to move When I was in China I was stunned by the strong manoeuvring ability of my opponent – a strong FM rated over 2400 Elo points. In this position, which arose in the exchange variation of the Slav, Jianjun Xie continued 18 ♘e2 planning ♘f4-h5 followed by ♕g4 and ♗e5. This is a logical plan in view of the fact that Black has slightly weakened his king’s position with ...h7-h6. As a matter of a fact one might say that an early exponent of this idea was Rubinstein who, in his game against Sämisch at Carlsbad 1929, induced a black knight to a5 by playing b4-b5 and then started play on the kingside in order to exploit the distant knight’s inability to get over to the defence. Nimzowitsch was very impressed by this idea and described it thoroughly in the book Carlsbad International Chess Tournament 1929. If you are not familiar with it, then it is well worth a look. The game continued 18...♖fc8 19 ♘f4 ♘ac4 20 ♕e2 ♘d6 21 ♘h5 ♘e8 22 ♕g4 ♗f8 23 ♗e5 with an active position for White. In this way White has involved all his pieces in the kingside operation. By reducing Black’s pieces to passivity, White can more easily choose the best plan. White’s dream, if he cannot break through on the kingside, is to exchange to an endgame where the bishop pair might be a decisive factor. 23...♕d8 24 ♕e2 ♘d7 25 ♗g3 ♘df6 26 ♘f4 Note that White will only agree to an exchange if there is positional gain in return. 26...♘d6 27 ♗h2 ♖xc1+ 28 ♖xc1 ♖c8 29 ♖d1! It would be a positional mistake to give away the c-file with 29 ♖xc8 ♕xc8. 29...♕a5 30 ♗b1 ♕b5 31 ♘d3 White’s knight endeavours to manoeuvre to c5 at the right moment. When you consider the manoeuvres the c3-knight has been capable of, you can only be impressed. There was no need to go via a4, which is so common in this opening, but rather go via the kingside and in the process activate White’s position much more than by traditional means. 31...♘de8 32 ♕e1 Black’s queen is misplaced on b5 and so it’s now time to go back with the rook to c1, since the queen is unable to defend the rook. 32...♘d6 33 ♖c1 ♖xc1 34 ♕xc1 ♕c4 35 ♕d1! ♕b5 36 ♕c2 ♕c4 37 ♘c5! Note that White’s queen manoeuvre to c2 has made possible a beneficial queen exchange. 37...♕xc2 38 ♗xc2 ♘c4 39 b3 ♘a5 40 b4 ♘c6 41 a3 ♗xc5 42 dxc5 White has achieved one of his goals and now has the bishop pair against two knights as well as a queenside pawn majority. White won a well played endgame after 87 moves. Good examples of playing over the whole board as well as showing how the principle of two weaknesses works can be seen in the following two positions. 107 Botvinnik – Zagoriansky Sverdlovsk 1943 White to move Black’s weakness on d5 is not enough for White to win the game so he needs to create a second weakness on the kingside. Here White can exploit the fact that Black has already weakened his kingside with ...h7-h6. If this wasn’t the case White would have had to manoeuvre to force Black to weaken himself on the kingside. But now White can open the g-file with g4-g5 and the second weakness will be the vulnerable black king. Botvinnik played 25 g4!. GM Baburin writes in Winning Pawn Structures – The Isolated Queen’s Pawn and its Derivatives (1998) that “Computers will not be able to come up with such a move for years to come, at least I hope not!” Twenty years later, he is to a certain extent correct because Komodo 11 suggests moves that mean that White should continue to tack about a little before embarking on his final plan. One of the moves that is suggested is 25 g3 with the idea of securing the kingsde with ♗g2 and ♔h2. Of course 25...♗xh3?? is then out of the question due to 26 g4 followed by ♕g3 or ♕h5 winning the bishop. 25...♕c6 26 g5 hxg5 A more active defence was 26...♕c2 with the idea 27 gxh6 ♕g6+ 28 ♔h2 ♕xh6 and the queen helps in defending the kingside. 27 ♕xg5 f6 28 ♕g6 ♗f7 29 ♕g3 29...f5? A serious positional mistake due to the dark-square weaknesses on g5 and e5. The pawn in itself also becomes a tactical weakness. Correct was 29...♗e6 deferring any specific course of action until White has shown his cards and revealed how he intends to attack on the kingside. White can attack with his rook(s), bishop or h-pawn. A bishop exchange on g4 is also possible. 30 ♕g5 ♕e6 31 ♔h1 ♕e5 32 ♖g1 ♖f8 33 ♕h6 White is hitting the tactical weakness on b6 while preparing to double on the h-file. 33...♖b8? A better defence was 33...♗e8 34 ♖h4 (or 34 ♕xb6 f4) 34...g6. 34 ♖h4 34 ♗h5 was also good, stripping Black of a most important defensive piece, and enabling him to exploit the resulting lightsquare weaknesses in Black’s camp. 34...♔f8 35 ♕h8+ ♗g8 36 ♖f4 ♖bb7 37 ♖g5 ♖f7 38 ♕h5 The f5-pawn is doomed. It’s a matter of taste how to win. 38 ♗h5 ♖f6 39 ♗g6 would also have been a decisive bishop manoeuvre to include the bishop in the attack and threaten 40 ♗h7. 38...♕a1+ 39 ♔g2 g6 40 ♕xg6 ♗h7 41 ♕d6+ ♖fe7 42 ♕d8+ This is good enough since Black resigned, but quicker and more precise was 42 ♗xd5 threatening mate on g8. Another game to study with this pawn structure and distribution of material is Petrosian – Yudovich, Moscow 1967. It’s not included in the present book but you can easily find it in a collection of Petrosian’s games. 108 Ivanchuk – Short Amsterdam 1994 White to move 13 a4! Preparations for f2-f4 can wait, since Black is not keen on playing ...f7-f5. White first wants to immobilise Black’s pawn majority on the queenside. If White manages to generate pressure on the queenside he can shift the attack from that side of the board to the king’s flank. 13 ♘f5 ♘e7 14 ♘xe7+?! (15 ♕g4!) 14...♕xe7 15 f4 exf4 16 ♖xf4 ♕e5 17 ♕c1 f5 is nothing special for White. 13...♖ad8?! 13...a5 was better. Black can then set up the pawn structure b6 and c5. In such a position White’s knight can manoeuvre between the squares c4 and d5. 14 a5! ♖d7 15 ♖a3! Black has a weakness on b7. If White manages to force Black’s pieces away from the kingside the pawn push to f4 will be more effective. This is the hidden point of the thirteenth move. 15...f6 16 g3 ♘e7 17 ♕e1! A very nice queen manouevre, preparing ♕b4 hitting the b7weakness. The time wasn’t ripe for 17 f4 due to 17...exf4 18 gxf4 f5 and Black has the initiative. 17...♖b8 17...♘c8 allows 18 f4. 18 ♕c3 ♔h8 19 ♕c5! White is threatening to penetrate with ♕a7 next move. 19...♘c8 20 f4! White has managed his deep positional idea very well and his play on the kingside has become more dangerous than it would have if he had played f4 at an earlier stage. 109-130: Pawn-Play and the Centre We first look at pawn-play on the wing and then pawn-play in the centre. The main idea is to use one or two pawns for positional gain. 109-114: Pawn-Play on the Wing We’ll look at four different situations. How to use a pawn to immobilise a knight, how to prevent an important break, how to play for your own breaks and how to prevent a battery from becoming operational. 109 Atkins – Capablanca London 1922 Black to move Black wants to exploit the bad position of the b3-knight. Capablanca played 14...a5!. By placing the pawns on a5 and b6, the b3 knight becomes passive and the weak b5-square can be defended by ...♘a7 if necessary. This is a set-up to remember! The alternative was 14...b6 15 ♖ac1 a5. 15 ♖ac1 15 ♘c5!? b6 16 ♘a4. 15...b6! 16 a4?! This weakens the b4-square. 16 a3 or 16 ♗e3 ♔d7 17 f3 ♘f5 18 ♗f2 was better. 16...♔d7 17 ♘c3 ♘a7! The knight on c3 is just as passive as the knight on a7. 18 ♔f1 ♘ec6 The problem with the king’s knight is solved. Normally it lands on f5. 19 ♔e2 ♖c8 20 ♗e1 ♗e7 21 ♘b1 f5! White must make an important choice. He can either allow Black to take more space with ...g5-g4 or he can open the game somewhat – but that would make the d4-pawn more vulnerable. 22 exf6!? ♗xf6 23 ♗c3 ♘b4 24 ♗d2 ♘ac6 White’s knight is still passive on b3 and the b4-square is weak. Black’s position is preferable. 110 Flohr – Spielmann Bled 1931 White to move 19 a4! This is an interesting counter-plan played by Flohr and Botvinnik to prevent Black from liberating himself by ...c5. Black cannot move the b5-pawn because of the battery on e2 and d3 and it’s difficult for him to do anything since ...c5 loses a pawn on b5. 19...g6 20 ♘d2 White’s knight obviously wants to go to c5 to block the c6-pawn and lock in the b7-bishop. Maybe it could also go to a5 in some positions so as to put pressure on c6, but then a doubling of rooks would probably be required. 20...e5 21 ♘e4 exd4 22 exd4 White’s strategy, which started with 19 a4, has been successful. His strong control of c5 means that the b7-bishop is practically dead – or else a “joke” to use an expression by Mikhalchisin. The most important factor when exploiting weak squares is to consider the opponent’s pieces which are trapped behind, in this case the b7-bishop. The strategy White has adopted here is always meaningful, since it effectively means playing with an extra piece. In this case the d3-bishop is decidedly stronger than its black counterpart. 22...♕b4 23 ♗b1! ♖e8 23...♕xa4? 24 ♘d6 ♖b8 25 ♕f3. 24 ♕d2 A nice technical move by Flohr. 24...♕xd2 25 ♖xd2 ♖b8 26 a5 Flohr was regarded as one of the best technicians in the world during this time and he managed to win this slightly better position after 55 moves. 111 Carlsen – Caruana Paris 2017 White to move In this blitz game Carlsen had the chance to play an unusual double pawn break after 16 b3 cxb3 17 c4!. In the game followed 17 ♘xb3 ♘e4 with a slight advantage to White. However, Carlsen still managed to play the important break c3-c4 after 18 ♖fc1 ♗c6 19 c4 and won after 33 moves. 112 Andersson – Brynell Skelleftea 1989 White to move Andersson played 25 b4! which is the only way to exploit his positional advantage. White plans the further break a4-a5 or an exchange on c5 followed by the exploitation of the open b-file with the queen. 25...b5? After strong moves weak replies come more easily. Other options were: a) 25...♖xe1+ 26 ♕xe1 ♕e8 although White has the better game after 27 ♕c3+ f6 28 bxc5 bxc5 (28...dxc5 29 ♕b3) 29 ♕a5 ♕e7 30 ♔f1 followed by ♕b6, ♘d2-b3-a5 with an attack on Black’s queenside pawns. b) 25...a5 26 bxc5 bxc5 27 ♕b2+ ♔g8 28 ♖xe8+ ♘xe8 29 ♕b6 and Black risks being suffocated. c) 25...cxb4 26 ♕b2+ ♔g8 27 ♕xb4 wins either the b6- or the d6-pawn. 26 bxc5 ♖xe1+ 27 ♕xe1 bxc4 28 ♕c3+ ♔g8 29 cxd6 ♘xd5 30 ♕xc4 ♕b7 30...♘b6 31 ♕c5 ♘d7 (31...♘xa4 32 ♕c7) 32 ♕c7 ♕e8 33 a5 and White wins. 31 ♘e5 ♘b6 32 ♕b3 Black resigned. 113 Fischer – Unzicker Olympiad, Siegen 1970 White to move 14 f5! This idea Fischer presumably borrowed from Lasker, who played a similar move against Capablanca in Saint Petersburg 1914. However, the first time Lasker played it was in his World Championship match against Steinitz in 1894. Then he tried it again against Janowski in another world title match in 1909. Lasker won all three games in which he adopted this idea. Lasker’s move was revolutionary since it went against the classical school. After all, White is left with a backward pawn on e4, a weak square on e5 and a devalued pawn majority on the kingside. Not even the computer is fond of the idea so what was Lasker doing? One of Fischer’s arguments was that if you want to get squares you’ve got to give squares. Nothing is for free in chess, not even squares! Here White concedes the e5-square but in exchange he gains control of e6 and at the same time confines the c8-bishop to passivity. The f6-pawn has been fixed and can be a future target with g4-g5. Presumably these fine points were taken into account by Lasker as well. 14...♕e7 15 ♗4 Fischer follows the same scheme that Lasker successfully employed against Capablanca. White wants to exchange the darksquared bishops and split the opponent’s bishop pair. Then White can prepare the future advance g4-g5. The game continued 15...♗xf4 16 ♖xf4 ♗d7 17 ♖e1 ♕c5 18 c3 ♖ae8 19 g4!? ♕d6 20 ♕g3 ♖e7 21 ♘f3 c5 22 e5! fxe5 23 ♖fe4 ♗c6 24 ♖xe5 ♖fe8 25 ♖xe7 ♖xe7 26 ♘e5 h6 27 h4 ♗d7?! The bishop was well placed on the long diagonal. Simpler and more logical was to play 27...♕d2, for example 28 ♕f2 ♕d5 29 ♘xc6 ♖xe1 30 ♕xe1 ♕xc6 with a most likely draw. 28 ♕f4! ♕f6 29 ♖e2 ♗c8 30 ♕c4+ ♔h7 31 ♘g6 ♖xe2 32 ♕xe2 32...♗d7? In this game Black doesn’t play well with his bishop and in fact the same mistake was made by Capablanca against Lasker. 32...♕d6!, with an eye on White’s exposed king, would have given Black good drawing chances. One plausible line is 33 ♕e8 ♕d1+ 34 ♔f2 ♕c2+ 35 ♕e2 ♗xf5! 36 ♘f8+ ♔g8 37 ♕xc2 ♗xc2 38 ♘e6 c4 etc. 33 ♕e7! White forces a transition to a winning knight versus bishop ending. 33...♕xe7 34 ♘xe7 g5 35 hxg5 35 fxg6+? ♔g7 is a draw. 35...hxg5 36 ♘d5 ♗c6 37 ♘xc7 ♗f3 38 ♘e8 ♔h6 39 ♘f6 ♔g7 40 ♔f2 ♗d1 41 ♘d7 c4 41...♗xg4 42 f6+. 42 ♔g3 Black resigned. 114 Sorokin – Belikov Elista 1995 White to move 11 c4!? White blocks the centre but at the same time Black’s battery becomes useless on the classical diagonal pointing at f2. Note that the kingside is devoid of defending pieces. The game continued: 11...d6 12 ♕g4 g6? This weakening pawn move wasn’t necessary. Correct was 12...00 13 cxb5 axb5 14 ♘f3 ♔h8 with mutual chances. 13 cxb5 axb5 14 ♘f3 h6 15 ♕h4 ♕d8? In a difficult situation Black makes another serious mistake. 16 ♗g5 ♕b6 17 ♗f6 ♖f8 18 ♕xh6 ♗a6 19 ♘g5 ♘e5 20 ♕g7 Black resigned. White was incredibly successful with his ugly-looking 12th move but it was a clever decision with regard to the situation in the centre. 115-130: Pawn-Play in the Centre 115 Lanka – I. Sokolov European Championship 1999 Black to move 18...d4! Black immobilises both knights with one pawn move. It’s true that Black’s bishop on b6 becomes passive but on the plus side the other bishop has possibilities of exerting pressure on the long lightsquared diagonal. The game continued: 19 ♘a3 ♕d5. Black has an incontestable battery on the long diagonal and threatens ...g5-g4 next move. 20 c4 ♕c6 21 h3 ♕g6! 22 ♘b5 ♖e6 23 ♗h2 ♘f4 24 ♗xf4 gxf4 25 ♘h2 h5 26 ♕d2 ♕f5 and Black won. The white knights are out of play for the rest of the game. Observe that sometimes it doesn’t matter if one bishop becomes bad or passive since Black has managed to make two knights ineffective! Chess is “give and take” and nothing is for free since everything comes at a cost – although in this case having his darksquared bishop confined to passivity was a small price to pay, considering what Black got in exchange. 116 Gligorić – Smyslov Candidates tournament, Yugoslavia 1959 Black to move 24...f6! Black plans 25...e5 which will restrict the movement of both the b2-bishop and the b3-knight. 25 g4 25 ♗d4 e5! 26 ♗xb6 ♖xd1+ 27 ♖xd1 axb6 25 ♔f1 ♖ac8 26 ♖xd8+ (26 ♖ac1 ♖xd1+ 27 ♖xd1 ♖c2 and Black wins.) 26...♖xd8 27 ♔e2 (27 ♖c1 ♖d3!) 27...♘xf2! 28 ♔xf2 ♖d3 and Black wins. 25...e5! 26 ♔f1 ♖ac8 27 ♖xd8+ ♖xd8 28 ♔e2 ♘xf2! 29 ♔xf2 ♖d3 30 ♘c5 ♗xc5 31 bxc5 ♖d2+ 32 ♔f3 ♖xb2 ...and Smyslov won the rook ending after a further 18 moves. 117 Romanishin – Polugaevsky USSR 1982 White to move In a rapid game Romanishin played the simulated pawn sacrifice: 14 e5!. The main idea is to take away the e5-square from Black’s pieces. while the next move secures the e4-square for White’s pieces. 14...fxe5? Black could have side-stepped his opponent’s conception by playing 14...♗c7, but after 15 f5 ♗f7 16 exf6 gxf6 17 ♗h6 White stands clearly better. 15 f5 This is the clever idea. Now that Black’s pawn has been lured to e5 the f-pawn passes it and the e4-square is in White’s possession. In this way White has gained an important square (space) in the centre and that is worth a pawn since Black cannot place a piece in the centre himself. White is actually positionally winning here. 15...♗d7 15...e4 would liberate Black’s game slightly but it’s not enough to change the sad state of Black’s position after 16 ♗g5! (16 ♗xe4 is also good but not 16 fxe6? e3! and White’s c1-bishop is practically dead.). 16 ♗g5 h6 16...e4 17 ♗xe4 ♗xg3 18 hxg3 and White wins or 16...♕c7 17 f6 gxf6 18 ♗xf6 with a strong attack. 17 ♗d5+ ♘xd5 17...♔h7 18 ♘g6!. 18 ♗xd8 ...and White won after 33 moves. 118 Nimzowitsch – Tarrasch Breslau 1925 White to move 20 f4 This move creates an outpost for the knight on d4. However, more precise was an immediate 20 ♘d4 ♖hd8 21 f4. Instead, after 20...e5 21 ♘b5 ♖hd8 22 ♖1c2 Black has problems with the defence of his a7-pawn, for example 22...b6 23 ♖a4 ♘c5 24 ♖xa7 ♖xa7 25 ♘xa7 ♖a8 26 ♘c6+ ♔d6 27 ♘b4 ♖a4 28 a3 and he is a pawn down. 20...♖hd8 Black could have disrupted White’s plans with 20...b5! when White has only a slight advantage. 21 ♘d4 f6 22 a4 White wants to prevent 22 b4 b5 but there was no need for this because of 23 ♘xb5 ♖d2+ 24 ♔f3 ♖xa2 25 ♘xa7 ♖dd2 (25...♘xb4? 26 ♘c8+) 26 b5 with a clear advantage. 22...e5? It was better to play 22...♔f7 23 b4 b6. Black has managed to contend with the outpost on d4 but now the e-pawn becomes weak. 23 fxe5 fxe5 24 ♘f3 ♔e6 25 b4 b6 ...and here Nimzowitsch missed the strongest continuation 26 ♖e4! with a clear advantage. 119 Gulko – Agdestein Reykjavik 1990 Black to move 16...d5! Gulko probably forgot that the en passant capture 17 exd6 wasn’t an option here due to the pin on the fifth rank. Such moves can easily be missed if one forgets about geometrical patterns, in this case the length of the fifth rank. After this mistake White’s game went downhill very quickly. 17 ♗g5 dxe4 18 ♗xe7 ♘xe7 19 ♖d7 ♕b4 20 a3 ♕xb2 21 ♖xe7 ♗d5 22 f4 c3 23 f5 c2 24 fxe6 c1♕ 25 ♖exf7 ♕d4+ 26 ♔h1 ♕xf1+ White resigned. 120 Andersson – Lahti Lidköping 1969 White to move As a youngster Ulf Andersson demonstrated another way to exploit a situation where there are two queens on the fifth rank. Black has tried to take advantage of the fact that the f5-pawn is pinned but White can turn this to his advantage by playing 18 e5!. This beautiful and at the same time logical breakthrough prepares the path for an attack on the black king. Note that such a move is in accordance with the principles laid out in Morphy’s games, see positions 3 and 4. 18...fxe5 19 fxe6 ♗xe6 20 ♗xe6 ♕xa4 21 ♖f1 Of course the presence of opposite coloured bishops is only to the advantage of the attacker, who can work on the light squares with his “extra” piece. This theme was also discussed in Bogoljubow – Alekhine, position 86. 21...♖c7 22 ♗b3 ♕c6 23 ♖xf7 ♔d8 24 ♕xe5 and White soon forced Black’s resignation. 121 Udovčić – Petrosian Zagreb 1965 Black to move Grandmasters Bronstein and Davies taught me that in such a situation you should always claim space in the centre because then you equalise immediately. So in this position the correct move is 3...e5!. The game Engqvist – Davies, Wrexham 1995, went 1 ♘f3 g6 2 g3 ♗g7 ...and now I played the inaccurate 3 ♗g2?! which was justly criticised after the game by Bronstein and Davies who recommended 3 d4!. Obviously Black responded 3...e5 with immediate equalisation. A recent example from the Swedish Championship at Ronneby 2018 was Von Bahr – Grandelius where 1 ♘f3 d5 2 g3 ♘d7!? was played. The obvious move now was 3 d4! but instead the less principled 3 c4 was played. 122 Kveinys – Engqvist Stockholm 1996 Black to move Here I played what was to my eyes the reasonable move 6...e6?! and after the game my opponent asked me why I didn’t play 6...e5 in one move instead of two. My argument was that I considered it inaccurate to place the bishop on g4 in the 2 d3 variation of the Caro-Kann, since it’s usually better placed on c8. After the immediate 6...e5 7 e4 dxe4 8 dxe4 ♗c5 Kveinys considered the position completely equal. White might continue 9 ♕e1 0-0 10 h3 and Black must make the important decision whether to exchange on f3 or retreat with either 10...♗h5 or 10...♗e6, but the position is definitely equal. The game continued 7 ♕e1 ♗c5 I regarded 7...♗e7 as too passive. 8 e4 dxe4 9 dxe4 e5 “Now you lose a tempo” Kveinys told me after the game, and he’s right, but what else could I do? 10 h3 White now enjoys a slight advantage whatever reply is made with the bishop. It’s not so strange. I’d actually lost a move and the balance was tipped in White’s favour, if only slightly – but an initiative is an initiative. The moral of these two little stories is that if you can claim space in one move then do it! I’ve already showed two situations where I’ve been criticised by experienced grandmasters for not doing it and I believe they are right, at least from the viewpoint of principles regarding how to relate to the centre and how to relate to activity versus passivity. 123 Vidmar – Nimzowitsch Carlsbad 1929 Black to move 8...cxd4! Nimzowitsch writes in the tournament book that this is an original concept which among other ideas is designed to prevent White from playing dxc5 followed by e4. It’s interesting that Komodo 10 also thinks this is the best continuation. 9 exd4 As a rule 9 cxd4 is unsuitable when White cannot place his knight on c3. 9...♘h5! The exploitation of the weak f4 square, accompanied by strong control of the b8-h2 diagonal, is the main idea of the early capture on d4. Another consideration is that now White cannot easily carry out the traditional plan with a knight outpost on e5. 10 ♘f1 h6 11 ♗d2 ♕c7 12 ♘g3 ♘f4 13 ♗xf4 ♗xf4 14 ♕d1 White has lost his bishop pair but in return he can play for the set-up 0-0, ♖e1, ♕e2 and ♘e5. 14...g6 Black prevents White from playing the knight to h5 but more accurate was 14...h5! at once. 15 0-0 h5 Nimzowitsch sets up the pawn wave formation on the kingside. 16 ♖e1 0-0 17 ♕e2 17 c4?! dxc4 18 ♗xc4 ♖d8 with pressure on the isolani is only in Black’s interest. 17...♔g7!? Nimzowitsch plans ...♖h8 followed by ...h4. The position is one of dynamic equilibrium. Black managed to win a well played game after 59 moves. The most important idea to consider is that pawn exchanges lead to weak squares in the position. In this case f4 was weakened and secured Black the bishop pair with a comfortable game. 124 Neikirch – Botvinnik Olympiad, Leipzig 1960 Black to move A correct understanding of the centre is probably the most important element in chess strategy. The centre is important in all phases of the game and not only the opening. 15...f6! Black could play 15...♗b7 to gain time and control the light squares in the centre, but it wouldn’t solve Black’s problem due to the strong e5-pawn. The drawback of 15...dxe5 is that following 16 fxe5 White would liberate his c1-bishop after 16...f6 17 ♗f4. Botvinnik’s solution is to attack White’s centre immediately before White can make use of it. 16 exd6 ♘xd6 17 ♖d3! ♘f5! Black’s knight is excellently placed on this outpost since it controls not only the centre but also observes squares on the kingside. 18 ♖a4? 18 ♗e3 was correct with a balanced position. 18...♕e8! 18...♗b7?? is answered by 19 ♖d7 ♖fd8 and now the tricky 20 ♖ad4!! ♘xd4 21 ♖xe7 wins. 19 ♘e4 b5 An immediate 19...♗b7! was stronger. 20 ♖a5? The retreat 20 ♖a1 was necessary. 20...♗b7 21 ♘d6 Not 21 ♘c5? ♗xg2+! 22 ♔xg2 ♕c6+. 21...♘xd6 22 ♖xd6 ♖d8 Black is threatening 23...♕c6!. 23 ♕d2 If 23 ♖xd8 ♕xd8 24 ♕e1 ♕d5 25 ♕e2 ♖d8 26 ♗e3 ♕d1+ 27 ♕xd1 ♖xd1+ 28 ♗g1 ♖d2. 23...♖xd6 24 ♕xd6 ♕d8! 25 ♕xe6+ ♖f7 26 ♕e1 ♖e7! White resigned. Note that all Black’s pieces exert pressure along the central files and the long diagonal. 125 Bernstein – Capablanca Moscow 1914 Black to move 15...c4! When this strong move was played by Capablanca it was initially regarded as a mistake by both critics and onlookers. The main reason was that it reduced Black’s influence on the d4 square and therefore made the advance of the d5 pawn more difficult. Also the d5-pawn itself had became weaker. However, Capablanca realised the value of the move and described it in detail in My Chess Career. White’s general plan was to put pressure on the hanging pawns with his rooks on c1 and d1 and Black would defend the pawns correspondingly by placing his rooks on d8 and c8. The e7-bishop defended the c5-pawn. Capablanca asked himself if it was possible to activate Black’s position and that is how he found 15...c4. The important point of Capablanca’s revolutionary move is that the e7-bishop becomes an active piece. Another important point is that the b2-pawn becomes a tactical weakness which Black can attack along the b-file. A third important issue is that if White blockades the d5-pawn by placing his knight on d4 the attack on the d-file is temporarily stopped. In addition to these arguments Nimzowitsch wrote in his classic My System that 15...c4 makes the hanging pawns relatively secure, because after this move Black has only one weakness instead of two. 16 ♖fd1 Capablanca recommended 16 e4 to simplify the game but Bernstein preferred to put pressure on the centre, since it was regarded as a good policy, and played for a win. 16...♖fd8 17 ♘d4 ♗b4! This is one of the points of Black’s 15th move. Black plans an exchange on c3. 18 b3 ♖ac8 19 bxc4 dxc4 Black’s d5-pawn has now been transformed to a passer so his overall strategy has certainly been successful. 20 ♖c2 ♗xc3 21 ♖xc3 ♘d5! 22 ♖c2 c3 23 ♖dc1 ♖c5 24 ♘b3 ♖c6 25 ♘d4 ♖c7 26 ♘b5 ♖c5 27 ♘xc3?? ♘xc3 28 ♖xc3 ♖xc3 29 ♖xc3 ♕b2! White resigned. A very important game to memorise since it illuminates several ideas regarding how to play with hanging pawns. These ideas are important to know whether you are White or Black. 126 Petrosian – Spassky Santa Monica 1966 White to move Here is another position with hanging pawns. The main drawback with hanging pawns is that they are grateful objects of attack. After every exchange their potential strength is reduced and in the endgame they are as a rule weak. The strength of hanging pawns lies in their mobility and ability to create sharp situations. It should be remembered that the semi-open e- and b-files have the potential for great activity and can be used as springboards to develop aggressive play. Here there is one disadvantage for Black, namely that the b7bishop stands in the way on the b-file. Compared with the previous position Black’s position is more passive and White is in a position to start a systematic siege of Black’s centre. 17 ♖c2 White plans an attack on the c5-pawn with ♖fc1 and ♘d2-b3. 17...♕d6 Black wants to place the queen in front of the bishop and exploit the b-file. 18 ♖fc1 ♕b6 19 ♘d2 White plans ♘b3 and ♗g4. 19...♖e8 19...♖c7 20 ♘b3 ♖ac8 21 ♗g4 c4 22 ♘d4 is a positional win according to Spassky but according to Komodo 11 it’s only slightly advantageous for White. Spassky probably wanted to say that it’s hard for Black to create any kind of activity and that spells suicide in this type of position where dynamic play is the key. 20 ♘b3 20...d4! Such a finesse is typical for the latent dynamic resources available in hanging pawns. 21 ♘xc5 ♘xc5 22 ♖xc5 dxe3 23 fxe3 ♖e7 23...♖ad8 24 ♕c3 ♖e7! was more precise according to Spassky. Now Black can choose whether to double on the e-file or the d-file. 24 ♖b5 Petrosian’s recommendation of 24 b4 ♖ae8 25 ♖1c3 gives Black the slightly better game after 25...♖d7. Spassky gives 24 ♗f1 ♖ae8 25 ♖e1 ♖d8 and Black has enough compensation for the pawn due to the pressure on the d-, e- and b-files. Also the long a8-h1 diagonal and the classical a7-g1 diagonal put White under positional pressure. 24...♕xe3+ 25 ♕xe3 ♖xe3 26 ♖xb7 ♖xe2 27 ♖cc7 ♖d8 28 ♖d7 ♖c8 29 ♖bc7 and a draw was agreed. 127 L. Karlsson – Engqvist Stockholm 1997 White to move Unusual but dynamic pawn structures are one of Swedish GM Lars Karlsson’s hallmarks. Here I was happy with the situation and thought I had equalised. I have stopped the dangerous break a4-a5 since then I would reply ...♘e5 with the double idea ...♘f3+ and ...♘c6. A pawn break in the centre with d3-d4 only results in hanging pawns and I didn’t mind putting pressure on them. For some reason I forgot about the most important pawn advance which was so close to Nimzowitsch’s spirit, namely... 17 f4! Now I suddenly had intricate positional problems to solve and felt I was being punished for having made some careless exchanges of a bunch of minor pieces earlier in the game. I also felt my game was rather uncomfortable due to potentially annoying aggressive moves like f4-f5 and ♘d5. I responded with the slightly desperate 17...f5. After the game I asked Karlsson what I should have played instead, but he gave me no answer. At home I checked with Fritz 6 which recommended the quiet developing move 17...♖ad8 with the verdict that the position is equal. Many years later I confirmed it with Komodo 11 which also recommended the rook move and with approximately the same evaluation as Fritz 6. My opponent played 18 e4!? but more precise was actually 18 a4! since it’s not possible to stop a4-a5 and a subsequent weakness on b6. It had no bearing on the practical result because after 39 moves I missed a mate in a dead lost position. He had a queen on f6 and a knight on f7 protected by a pawn on e6 and I overlooked 39 ♕h8 mate. Obviously I had forgotten that the knight controlled h8. After this crush he told me I’d been completely out of it during the game. Maybe so, but other reasons were presumably that I wasn’t used to this kind of pawn structure nor the protected knight on f7 which controlled the all-important corner square. 128 Stein – Rodriguez Havana 1968 White to move Stein played 19 e4! ♘c7 20 d4. This idea has been known ever since the game Réti – Rubinstein, Carlsbad 1923. It means that White’s two pawn moves express a concept known as “the hypermodern paradox”. This was coined by Reuben Fine in his book The World’s Great Chess Games and the expression is defined like this: “If you occupy the center immediately, the pawns soon become weakened, and the opponent will then be able to occupy it. However, this means that it is not occupation that is bad, but in many cases immediate occupation.” 20...exd4 21 ♘xd4 Paradoxically it’s White who has a pawn in the centre but three moves ago it was Black. Here the e4-pawn is strong since White has active pieces to support his centre. 21...♗xd4 22 ♗xd4 ♘b5 23 ♗b2 and Stein had a slight advantage and won convincingly after 47 moves. 129 Kasparov – Karpov World Championship match, New York 1990 White to move Kasparov played 19 f3!. This strong and solid move was actually discovered by Timoschenko with the principal idea of just maintaining the nice classical centre and avoiding any change in its structure, otherwise it would be easier for Black to attack it. However, one important point of this deceptive and simple move, is that the three black pieces on a6, b7 and e8 will have no work to do. The rook on a6 is out of play when White is solid in the centre. As a matter of a fact the three other moves Kasparov analyses in his famous white cover books give Black an easier game since the central structure will then be easier to crack. Black’s reward after breaking up the centre is his more active pieces. Let’s look at Black’s activity in the following variations: a) 19 e5?! dxe5 20 dxe5 ♕d5 21 ♘hf3 ♖ae6! 22 ♖ae3 ♗c5 23 exf6 ♗xe3 24 fxe3 ♖xe3; b) 19 ♘g4 ♘xg4 20 hxg4 (20 ♕xg4 c5! 21 dxc5 dxc5 22 e5 ♕d4 23 ♕g3 ♖ae6 and Black equalised in Hjartarson – Karpov, Seattle 1989.) 20...♗c8; c) 19 f4 d5 20 e5 ♘e4 21 ♘g4 c5! 22 ♘xe4 dxe4 23 dxc5 ♗xc5+ 24 ♗e3 ♗f8 and Black had no problems in the game Ivanchuk – Karpov, Linares 1989. 19...♕d7?! Presumably 19...c5!? is the best move, although the rook on a6 is cut off after 20 d5. However, Black has the strong continuation 20...♕e7 threatening the d5-pawn. Possibly Kasparov would have continued 21 ♖f1!? ♗g7 22 ♘c4 ♘h5 23 ♘g4!? ♘g3 24 ♘xh6+ ♔h7 25 ♘g4 ♘xf1 26 ♕xf1 ♖aa8 27 f4 with some initative to compensate for the exchange sacrifice. The reason Karpov avoided this was probably because he wanted to avoid Kasparov’s preparation and anyway he normally refrained from such Benoni formations. 20 ♘c4 ♕b5 21 ♖c3! ♗c8 From now on Karpov plays in a passive manner which shows what a psychological effect White’s 19th move had. Black’s position deteriorates while White steadily improves his chances. 22 ♗e3 ♔h7?! 23 ♕c1 c6? 24 ♘g4 ♘g8? 25 ♗xh6! ♗xh6 26 ♘xh6 ♘xh6 27 ♘xd6 ♕b6 28 ♘xe8 ♕xd4+ 29 ♔h1 ♕d8 30 ♖d1! Most probably Kasparov had this position in mind as a steppingstone when he started the combination at move 25. 30...♕xe8 31 ♕g5 Black has completely lost the coordination of his pieces. 31...♖a7 32 ♖d8 ♕e6 33 f4 ♗a6 34 f5! ♕e7 35 ♕d2 ♕e5 36 ♕f2! ♕e7 37 ♕d4 ♘g8 38 e5 ♘d5 39 fxg6+ fxg6 40 ♖xc6 ♕xd8 41 ♕xa7+ ♘de7 42 ♖xa6 ♕d1+ 43 ♕g1 ♕d2 44 ♕f1 Black resigned. 130 Tarrasch – Noa Hamburg 1885 White to move This position arose in the Steinitz variation of the Classical French Defence after 1 e4 e6 2 d4 d5 3 ♘c3 ♘f6 4 e5 ♘fd7. Now the question is whether White should go for a short pawn chain or a long one. 5 ♘ce2 Réti writes in his ground-breaking Masters of the Chess Board that it’s a mistake to place the pawns on c3 and f4 since this plan is too ambitious. In principle White wants to suffocate Black but as Steinitz in his insightful way taught us you must always play in accordance with the demands of the position. From that perspective it’s right to be happy with a shorter pawn chain by playing an immediate 5 f4 and answering 5...c5 with 6 dxc5. White should be satisfied with the spearhead at e5 and give up any attempts to defend d4 as well. Black hasn’t made any mistake and according to the principle of accumulation one should just gather small advantages, such as the space White gains from having his pawns on e5 and f4. According to the principle of proportion, one of Lasker’s famous theories, one should make a move proportionate to the advantage one has in the position. Here White’s advantage is not big enough to be able to maintain the whole pawn chain with the support of c2-c3 and f2-f4. 5...c5 6 c3 ♘c6 7 f4 Now the best continuation was to keep the tension in the centre by 7...♕b6 followed by ...f6, ...♗e7 and ...0-0 with a good game for Black. 131-147: Play on the Wings The last part in this section deals with play on the wings. First the minority attack is discussed, followed by the queenside and kingside attack. We will mainly deal with different methods of managing the pawns rather than pieces, with the exception of position 145 which deals with one of Fischer’s favourite themes: “empty squares”. 131-135: The Minority Attack 131 Karpov – Ljubojević Linares 1989 White to move The minority attack is a common positional theme in modern chess. This type of structure often arises in the exchange variation of the Queen’s Gambit, as seen here, and the exchange variation of the Caro-Kann with the black pieces. White strives for b2-b4-b5 to weaken Black’s queenside formation, in particular the c6-pawn. If Black captures the b5-pawn the d5-pawn will become a weakness and if Black allows White to capture on c6 then the c6-pawn will be a permanent weakness. 13 b4 a6 It’s useful to get rid of the black a-pawn since it risks becoming a weakness when White carries out his minority attack. 14 a4 ♗f5 15 ♘e5 ♖ad8 16 ♖fc1 ♘g6 17 ♗xe4 ♗xe4 18 ♘xe4 dxe4?! 17...♘xe5 was better, in order to force 18 ♘d2. 19 ♘xg6 hxg6 20 b5 cxb5 21 axb5 ♖d6 22 bxa6 bxa6 23 ♕a4 ♕d7 24 ♕xd7 ♖xd7 25 ♖c5 ♖a7 26 ♖a5 ♔f8 27 ♖b6 ♖ea8 28 h4 Karpov eventually won by applying pressure against Black’s pawn weaknesses on a6 and e4. His timing of the b4-b5 break was impressive. He waited patiently until all the minor pieces had been exchanged to minimise Black’s counterplay on the kingside. Karpov’s studious technique made it only a matter of time before Black’s weaknesses were exploited one after another in an endgame with only major pieces on the board. 132 L.B. Hansen – C. Hartman Gausdal 1991 White to move 16 ♔a1!? White is preparing a minority attack despite the fact that his own king has castled on the queenside! 16...h5 17 b4 A risky advance, but a useful idea to remember. 17...h4 17... ♕xb4? 18 ♘c5 ♕a3 19 ♖b1 ♗c8 (19... b6 20 ♖b4! ♕a5 21 ♖eb1) 20 ♘xb7! ♗xb7 21 ♕xc6 ♖d7 22 ♗a6 ♕e7 23 ♗xb7 ♖xb7 24 ♖xb7+ ♕xb7 25 ♖b1. 18 ♘f1! The seemingly passive f1-knight has important defensive duties since it protects the h2-pawn and if necessary can also cover the c4square by ♘d2. 18...♗f5 19 ♘c5 ♗xd3 20 ♘xd3 ♘f5 21 ♕b3 ♘d6 22 a4 ♖c8 23 ♖c1 ♘c4 One of the main drawbacks of the minority attack is that if Black manages to exchange the light-squared bishops the c4-square might become available for a black knight. However, in the present position White has a strong central thrust. 24 e4! ♕d8 25 e5 ♘e4 A complicated position has arisen where White managed to score a full point after 84 moves. 133 Mozny – Dumirache Odorheiu Secuiesc 1995 White to move 13 a4?! This routine move gives Black counterplay along the f8-a3 diagonal. Correct was 13 ♘a4!. It’s important to manoeuvre the knight to c5 before playing a4. On c5 the knight would have helped to plug the f8-a3 diagonal. 13...♗g4 14 ♘d2 ♗e7 Now came the natural reply 15 b5 (15 ♖ab1 was more circumspect.) but it was effectively met by 15...a5!. This idea originates from one of the great experts of the Queen’s Gambit, Boris Spassky. The point is that White cannot play the knight to c5 via a4 and besides this Black has got a nice outpost on b4 for his bishop. White must have felt he had been fooled in this position – so remember not to play a2-a4 without thinking! 134 Korchnoi – Karpov World Championship match, Baguio City 1978 White to move 19 a5! White wants to fix the black queenside pawns and so prepares the manoeuvre ♘c3-a4-c5. Note that a traditional minority attack with 19 b5? is not possible here due to 19 ... cxb5 20 ♘xb5 a6 21 ♘a7 ♖c2 when Black has a lot of counterplay. 19...a6 20 ♘a4 An obvious manoeuvre when Black has a static weakness on b7. 20...♗f8 21 ♘c5 ♖e7 21...♗xc5?! 22 bxc5 followed by doubling on the b-file forces Black to defend the b7-pawn with his rooks, thereby making them look silly. 22 ♔f1 ♘e8 23 ♔e2 ♘d6 24 ♔d3 Note that White’s king is more active than its counterpart. 24...♖ce8 25 ♖e1 g6 Here the players in Karpov’s camp thought Black had the advantage but they were far from being right. The position is balanced. 26 ♖e2 f6 27 ♖ae1 ♗h6 28 ♘db3 ♗f8 29 ♘d2 ♗h6 30 h3 ♔f7 31 g4 The moment has come to provoke an additional weakness on the kingside. 31...♗f8 32 f3 ♖d8 33 ♘db3 ♘b5 34 ♖f1 ♗h6 35 f4! ♗f8 36 ♘d2 ♘d6 37 ♖fe1 h6 38 ♖f1 ♖b8 39 ♖a1 ♖be8 40 ♖ae1 ♖b8 41 e4! dxe4+ 42 ♘dxe4 ♘b5 43 ♘c3 ♖xe2 44 ♖xe2 ♗xc5 (44...♘c7 45 f5) 45 bxc5 ♖d8 46 ♘xb5 axb5 47 f5 gxf5 48 gxf5 The kingside pawns have now also been fixed. 48...♖g8 This interesting rook ending will be discussed in the endgame section, position 259. 135 Monticelli – Capablanca Barcelona 1929 Black to move This type of position is slightly more comfortable for Black due to the fact that he has the possibility ...a7-a5-a4. The equivalent pawn push e2-e4-e5 is more difficult to achieve. 10...♘e4! It will be easier to carry out the pawn push to a4 after the exchange of knights. 11 ♕c1 ♘xc3 12 ♗xc3 ♗xc3 13 ♕xc3 a5! Black’s long-term plan with the minority attack (one against two) is to weaken the b3-pawn. 14 ♖fd1 ♘d7 15 ♕e3 ♔g7 Black must play actively with his king and not allow the white queen to invade the h6-square. 16 h3 ♘f6 17 g4 h6 18 ♘e1 ♕c7 19 ♘d3 An interesting knight manoeuvre was 19 ♘c2 a4 20 ♘a3 when he could counteract the pressure on the b-file by placing his knight on b5. 19...♗xg2 20 ♔xg2 a4! Black has achieved the first step in his plan. 21 ♘f4 ♖fb8 22 f3? axb3 23 axb3 ♖xa1 24 ♖xa1 24...e5! White probably missed this move since the knight cannot go back to d3 due to the pressure on the b3-pawn. 25 ♘d5 ♘xd5 26 cxd5 ♖b4! White cannot hold on to the weak pawns on b3 and d5 and therefore has a lost position. 27 ♕d3 ♕b7 28 ♕a6 ♕xd5 29 ♕c8 ♕b7! 30 ♕d8 ♕b8 31 ♕d7 ♖b7 32 ♕c6 ♖xb3 33 ♖a6 ♖b2 White resigned. White made some errors in the process of dealing with the minority attack. It required exact play to stand up against it and that is not always easy in a practical game. An important plan to remember! 136-137: Queenside Attack Next follow two model examples of how to attack the queenside when there is no minority or majority. It’s rather a matter of qualitative pawns since it’s obvious that a quantitative majority in itself is a signal for a duly prepared assault, especially when the attacker also controls an open central file. 136 Botvinnik – Boleslavsky Moscow 1939 White to move 22 b4! Black’s b6-knight invites the advance of the pawns on the queenside. This means that White’s pawns will represent a qualitative pawn majority, to borrow an expression from Nimzowitsch, rather than quantitative. White’s superiority on the queenside lies in the fact that his pawns are more active than Black’s and create more space. 22...♗e6 22...f5? 23 exf5 (23 ♘h4 with the double threat on f5 and g6 is also decisive.) 23...gxf5 24 ♘h4 and White wins. 24 ♖xd7 ♖xd7 25 ♘xf5 with a dangerous attack for White according to Botvinnik – although Komodo 10 doesn’t agree. After the sequence 25...♖xd1+ 26 ♕xd1 ♕d7 27 ♕b1 ♘d5 White has enough compensation for the exchange but no win. 23 ♗b3 ♖xd2 24 ♕xd2 ♗xb3 25 axb3 White plans c4-c5 followed by an invasion on the seventh rank. 25...♕e6 26 c4 ♗f6 27 c5 ♘c8 27...♖d8 28 ♕xd8+ ♗xd8 29 ♖xd8+ ♔g7 30 cxb6 ♕xb3 31 ♖d7 is hopeless since the b7-pawn is lost. 28 ♕d7 and the game is decided due to the invasion by the queen on the seventh rank. White won after 46 moves. 137 Mihalchisin – Djurić Copenhagen 1979 White to move 11 a4 The idea is to start an attack on the queenside by playing a5-a6 followed b4-b5 and undermining the light squares in Black’s position. This plan made a deep impression on GM Adrian Mikhalchisin (once upon a time Karpov’s trainer) when GM Yossif Dorfman (Kasparov’s erstwhile assistant) introduced it to him. Remember that it’s useless to attack on the kingside or in the centre. 11 e4?! weakens the d4-square and is strongly answered by 11...e5! when it’s White who must fight for a draw. 11...a5 Now it looks like White is weak on b4 but since the black knight has already developed to d7 it’s difficult to exploit it. White continues 12 b3 with the intended plan of ♗a3, ♖ab1 and b4 when the b-file will be opened. 12...♖e8 13 ♗a3 ♕b6 14 ♖ab1 Mikhalchisin plays according to Dorfman’s instructions. 14...♖ad8?! 14...♗f8 would have made it more difficult for White to play b2b4. 15 b4 axb4 16 ♗xb4 ♕a7 With his rooks on the d- and efiles, Black plans to open the centre with ...dxc4 followed by ...e6-e5. So White prevents this with... 17 cxd5 exd5 17...cxd5? 18 ♘b5 is obviously a “knightmare” for Black. 18 ♕e2 White wants to bring the queen over to the queenside. 18...♖e6 19 a5 Black has a weakness on b7 and White can attack it by placing the queen on a2 and follow it up with a knight manoeuvre to c5 and then a doubling of the rooks on the b-file will finish the job. You can only say: “Poor b7-pawn!” 138-147: Kingside Attack The first thing to learn is how to attack the opponent’s king when the centre is closed or secure. You are recommended to study several Steinitz games in depth. Here only a few positions are presented but some other important games are mentioned in the annotations. 138 Dubois – Steinitz London 1862 Black to move Steinitz played 7...g5 Another good move was 7...♘a5 to secure the bishop pair. 8 ♗g3 h5 Preparing a pawn storm against White’s king. This game is regarded as the prototype of how to attack the king with a pawn storm. The main reason such an attack has good prospects of success is because it’s difficult for the defender to counter with a breakthrough in the centre since it is of a closed character. The game continued 9 h4. 9 ♘xg5 is best met by 9...h4 10 ♘xf7 ♕e7 11 ♘xh8 hxg3 12 hxg3 ♕h7 with the threat of 13...♘g4. 9...g4! In the actual game Steinitz wanted to castle queenside after 9...♗g4 10 c3 ♕d7 but this is shown to be inferior if we check the variations with the computer. 10 ♘g5 0-0! It may seem paradoxical to castle kingside but the king is actually quite safe on g8. At the same time the g3-bishop bites on granite. It’s only in conjunction with c3 and d4 that the g3-bishop can come alive but that isn’t easy to achieve. 139 Aronian – Kramnik Candidates tournament, Berlin 2018 Black to move This position arose after 1 e4 e5 2 ♘f3 ♘c6 3 ♗b5 ♘f6 4 d3 ♗c5 5 ♗xc6 dxc6 6 0-0 ♕e7 7 h3. Kramnik improved on Steinitz’s prototype, where ...h7-h6 and ...g7-g5 was played, by making the clever move 7...♖g8!. The main idea is to play ...g7-g5 without wasting a tempo on ...h7-h6. 8 ♔h1 ♘h5 8...h6 would have been according to the prototype. 9 c3 g5! This is the idea. The game continued 10 ♘xe5 g4 11 d4 ♗d6 12 g3 ♗xe5 A good alternative was 12...♗e6. 13 dxe5 ♕xe5 14 ♕d4 ♕e7 14...♕xd4 15 cxd4 gxh3 was playable but understandably Kramnik wanted to keep the queen in play. 15 h4 c5 16 ♕c4 ♗e6 17 ♕b5+ c6 18 ♕a4? Correct was 18 ♕d3 ♖d8 19 ♕e3 ♗c4 20 ♖g1. 18...f5 19 ♗g5 ♖xg5 20 hxg5 f4 21 ♕d1 ♖d8 22 ♕c1 fxg3 23 ♘a3 ♖d3 24 ♖d1 ♗d5! 25 f3 gxf3 26 exd5 ♕e2 27 ♖e1 g2+ White resigned. A very inspirational game by Kramnik. 140 Steinitz – Lasker World Championship match, New York 1894 White to move 11 h4! White launches a direct attack since the black king has already taken up permanent residence on the kingside, while his own central position is solid. This procedure could justly be called the Steinitz plan. The prototype is Dubois – Steinitz, London 1862, but Steinitz – Chigorin, Game 4, World Championship match 1892, and Steinitz – Blackburne, London 1876, are other famous games you can study to obtain a more complete picture of this important concept. The Russian theoretician Suetin has made the very interesting statement that this kind of attacking move marks the transition from the opening to the middlegame. 11...♕c7 12 ♘g5 White plans h4-h5 with his next move. 12...d5 13 f3! Steinitz maintains his solidity in the centre. 13...♖ad8 14 g4! The main idea when in possession of a solid centre is to create conditions for a successful attack against the opponent’s castled king. 14...dxe4 15 fxe4 h6 16 ♕f3! ♗e8 17 ♗c2 ♘d7 Here White carried out a very interesting manoeuvre. If you have forgotten it you can go back to position 72. 141 Lasker – Capablanca Moscow 1935 White to move White is very solid in the centre with pawns on d4 and c3. His central position hasn’t been under attack with an early ...c7-c5 and in accordance with Steinitz’s principles this means that White will not be disturbed when launching a dangerous attack on the kingside. Therefore Lasker duly played: 19 h4!. White has three ideas in mind. He wants to play h4-h5 at the right moment and place his queen on a dark square, exchange the dark-squared bishops and then advance his pawn to h6 and if possible the queen to f6. A second possibility is to exchange on g6 and exploit the g5-square and the h-file. A third possibility is to prepare a sacrifice on g6 which will have been weakened after the pawn exchange. 19...♕d8 20 h5 ♕g5 21 ♗xg7 ♔xg7 22 ♖e5 ♕e7 23 ♖de1 ♖g8 24 ♕c1! It’s necessary to place the queen on a dark square to be able to break through with it on the kingside. 24...♖ad8 25 ♖1e3 ♗c8 26 ♖h3 ♔f8 27 ♕h6+ ♖g7 28 hxg6 hxg6 29 ♗xg6! White destroys Black’s pawn cover with this thematic idea which is made possible when the attacker’s pieces are as active as they are. 29...♕f6 29...fxg6!? 30 ♕h8+ ♔f7 (30..♖g8 31 ♖f3+) 31 ♖f3+. 30 ♖g5! ♔e7 31 ♖f3 ♕xf3 32 gxf3 and White won. 142 Polugaevsky – Ftáčnik Olympiad, Lucerne 1982 Black to move 17...h5! Black plans to profit from the advance h2-h3 by weakening White’s kingside further with ...h5-h4 and ...hxg3. White cannot really avoid this by playing the g3-pawn to g4, because Black can reply ...g7-g5 and ...♘g6-f4 with a positional advantage. This kind of kingside attack is exceptionally dangerous, since all pieces can be deployed in the attack as there have not yet been any exchanges. 18 f4 ♘g6 19 ♘f3?! 19 h4? can obviously be met by 19...♘g4. 19 f5 was better although Black stands well after 19...♘e5 as he has strong control of the dark squares. White’s g2-bishop is not so good either. 19...d5! 20 cxd5? 20 e5 is better although 20...♘e4 will highlight White’s weaknesses on the kingside after 21 ♘xe4 dxe4 22 ♘d4 h4 etc. 20...h4! 21 ♘xh4 21 dxe6 ♗c5+ 22 ♔h1 ♘h5 and Black wins. 21...♘xh4 22 gxh4 ♕xf4 23 dxe6 fxe6 24 e5 ♗c5+ 25 ♔h1 ♘h5 26 ♕xh5 ♕g3 27 ♘d5 If 27 ♕e2 ♕xh3 mate. 27...♖xd5 28 ♖f1 28...♕xg2+! 29 ♔xg2 ♖d2+ White resigned. If White had played on a few moves he would have been mated after 30 ♔g3 ♖g2+ 31 ♔f4 ♖f8+ 32 ♕f7+ ♖xf7 mate. The last reserve in Black’s huge army executes the final blow. It was because of games like this that players started to exchange pieces in the Hedgehog so as to disarm its huge attacking potential. 143 Zaragatski – Kovalev European Cup 2013 Black to move 10...h5!? Such a move can be worth playing even though the rook is not behind the pawn. One of the ideas is to secure the g4-square for Black’s minor pieces. 11 ♘c3 On 11 h3 play might continue 11...h4 12 g4 ♘xg4 13 hxg4 ♗xg4. 11...h4 12 ♘a4 hxg3 13 hxg3 ♘e6 14 c5 ♘g4 This idea is important to remember both for the attacker and the defender. When the h-pawns have disappeared from the board it’s difficult for White to drive away a black knight on g4 and in addition Black can follow up by activating his queen to h5 via g5. Black had the advantage and won after 15 cxd6 cxd6 16 ♖c1 ♗d7 17 ♗xb7 ♖b8 18 ♗g2 ♕g5! 19 ♘f3 ♕h5 20 ♘c3 f5 21 ♖e1 f4! 22 ♘d5 fxg3 23 fxg3 ♖xb5 24 ♘e7+ ♔f7 25 ♗a3 ♔xe7 26 ♕a4 ♗h6 White resigned. Don’t forget this clever idea removing one file of the board. It’s reminiscent of position 98 where one rank was removed. 144 Tartakower – Lasker Saint Petersburg 1909 Black to move 42...g5! Lasker follows the well-known principle that when the attacker has a superiority in the centre the game can be decided with an attack on the flank. Another plan suggested by Komodo 11 was the advance 42...h5 followed by ...h5-h4. In such a situation White must be careful of both ...h4-h3 and ...hxg3, weakening the kingside. Lasker’s move is more logical and human since it obtains strong pressure against f4. 43 ♕a2 The best defensive move in a lost position. 43 fxg5?! ♘e5 44 ♗e1 (44 ♗c5 loses quickly after 44...♖d3 45 ♖c3 ♘f3+ 46 ♖xf3 ♕xf3 47 ♖xd3 ♖xd3 and White must sacrifice the queen to avoid mate. Also after 44 ♕e3 ♘f3+ 45 ♔h1 ♕a8! Black has the threat of an unstoppable discovered check which leads to mate or material gain.) 44...♖d1! and the e1-bishop is hanging. 43...gxf4 44 ♖e2 ♕g6 45 ♕c2 ♔h7 46 ♕c3 ♖g8 47 ♔h1 ♕h5 48 ♖d2 fxg3 49 ♗xg3 ♖xg3 50 ♕c6 ♘e5 51 ♕e4+ ♔g8 52 ♖df2 ♖g5 53 ♖c2 ♖d1 White resigned. 145 Fischer – Miagmasuren Sousse 1967 White to move Fischer spent most time (15 minutes) during the whole game on the apparently harmless move 13 a3. He obviously wanted to prevent ...a4-a3 which would have forced White to play b2-b3 with resulting weaknesses on c3 and d4. The drawback of the move is that White loses a tempo in advancing his attack on the kingside. The revolutionary idea is that it went against established theory since the time of Steinitz and the 1951 World Championship match that you shouldn’t weaken yourself on the same side where you were being attacked. 13...bxa3 14 bxa3 ♘a5 15 ♘e3 Note that White doesn’t aim for h5-h6 in this position since he wants to keep these squares available (empty squares) for moves such as ♘g5 or ♕h5. It is typical of Fischer to strive for manoeuvring space for his pieces and not let a pawn get in the way. 15...♗a6 16 ♗h3 d4 17 ♘f1! Stronger than the routine move 17 ♘g4, since the knight can head for e4 via d2. Another drawback in playing the knight to g4 is that it would temporarily obstruct the queen along the d1-h5 diagonal which again would not be in keeping with Fischer’s style. 17...♘b6 18 ♘g5 ♘d5 19 ♗d2 ♗xg5 20 ♗xg5 ♕d7 21 ♕h5 White exploits to the maximum the manoeuvring space available for his queen. 21...♖fc8 22 ♘d2 ♘c3 23 ♗f6! ♕e8 24 ♘e4 g6 25 ♕g5 ♘xe4 26 ♖xe4 c4 27 h5 cxd3 28 ♖h4! ♖a7 29 ♗g2! dxc2 30 ♕h6 ♕f8 31 ♕xh7+!! Black resigned. This beautiful attacking game would never have become a reality were it not for Fischer’s appreciation of “empty squares”, which means not allowing pawns to get in the way of pieces! 146 Spassky – Petrosian World Championship, Moscow 1966 Black to move 19...a6! A classical approach to prevent opening of lines. If White advances the pawn to a5 Black answers ...b5 and vice versa. Petrosian now had a free hand on the kingside and won one of his most beautiful games ever. A slightly more complex example of the same idea can be seen in the next position. 147 Agdestein – Olafsson Nordic Championship 1985 Black to move 14...g6! This move is reminiscent of ...a6 which was played in the previous example. In that game Petrosian prevented the opening of lines on the queenside and here Black does the same, but on the kingside. Black plans ...♗g7 when it’s possible to meet h4-h5 with ...g6-g5 and g4-g5 with ...h6-h5. 15 h5 15 g5?! opens lines but is more risky for White than for Black after 15...hxg5 16 hxg5 (16 h5? ♔g7 17 hxg6 fxg6 is clearly better for Black,) 16...♗xg5+ 17 ♔b1 ♕f6 due to the strong control of the black squares. The game ended peacefully after: 15...g5 16 ♕f5 ♘b4 17 ♘e5 ♗c8 18 ♕f3 ♗xe5 19 dxe5 ♕c7 20 ♔b1 ♕xe5 21 ♗g2 ♗a6 22 a3 d4 23 axb4 dxc3 24 ♖he1 c2+ 25 ♔xc2 ♕c7+ 26 ♔b1 ♖ad8 27 ♕f6 ♗d3+ 28 ♖xd3 ♖xd3 29 ♕xh6 ♖c8 30 ♕xg5+ ♔f8 31 ♕h6+ ♔g8 32 ♕g5+ ♔f8 33 ♕h6+ and a draw was agreed. 148-150: Psychological and Pragmatic moves We now turn to a very interesting part of the book where we touch on the human aspect of chess and how we can deal with different situations, both as a defender (148) and as an attacker (149-150). Psychological chess has many dimensions and you will have some comments now and then about these issues when discussing other positions. In this section we are focusing on pragmatism and how it can affect the play psychologically. 148 Averbakh – Spassky USSR Championship 1956 Black to move 16...♘c6!? Brilliant ideas and a large dose of originality was Spassky’s signum. This move is regarded as one of the most incredible moves ever played. Presumably Spassky reasoned like this: Black has a very cramped position without any counterplay. Normal moves fail to improve his position and he would then have been slowly squeezed to death and the game decided on the h-file. Positionally he’s already lost so what to do? The only way is to be optimistic and play psychologically. After 17 dxc6 bxc6 he has a semi-open b-file and the possibility of getting a knight on d4 after the manoeuvre ...♘e6-d4. Spassky’s idea was so strange that the audience thought that something was wrong with the demonstration board. What is even more incredible is that the great technician Averbakh failed to exploit his material advantage and the game ended in a draw after 73 moves. If you are a professional player I recommend you play through the rest of the game in your head without a board and answer the question why the end position is a draw. The rest of you can do the same but with a board. The rest of the moves were: 18 ♘h4 ♕e8 19 hxg6 hxg6 20 ♕g4 ♖b8 21 ♘d1 ♘e6 22 ♖a3 ♘d4 23 ♖ah3 ♕f7 24 ♗c3 ♖fe8 25 ♖3h2 ♕xc4 26 ♘xg6 ♖e6 27 ♗xd4 ♖xg6 28 ♕f5 ♕e6 29 ♕xe6+ ♖xe6 30 ♗c3 d5 31 f3 ♖b3 32 ♖h3 c4 33 ♔d2 ♖g6 34 ♖g1 d4 35 ♗a5 ♗f8 36 ♖g4 ♖d6 37 ♔c2 ♖d7 38 g6 ♖db7 39 ♗e1 c5 40 ♖gh4 ♗g7 41 ♗a5 c3 42 bxc3 ♖a3 43 cxd4 exd4 44 ♖xf4 ♖a2+ 45 ♔d3 ♖b1 46 ♖h1 ♖xa4 47 ♔c2 ♖b5 48 e5 d3+ 49 ♔xd3 ♖xf4 50 ♗c3 ♖xf3+ 51 ♔e4 ♖g3 52 ♔f4 ♖xg6 53 ♘e3 ♖b8 54 ♘f5 ♖f8 55 ♖h5 ♖e8 56 ♔e4 ♖g1 57 ♖h3 ♗f8 58 ♔d5 ♖d1+ 59 ♔e4 ♖c1 60 ♔d5 ♖d1+ 61 ♔e4 ♖d7 62 ♘h6+ ♗xh6 63 ♖xh6 ♖h7 64 ♖g6+ ♔f7 65 ♖f6+ ♔e7 66 ♖c6 ♔d7 67 ♖xc5 ♖h6 68 ♔d5 ♖b6 69 ♗a5 ♖b5 70 ♖xb5 axb5 71 e6+ ♖xe6 72 ♔c5 ♖e5+ 73 ♔b6 and a draw was agreed. 149 Karpov – Spassky Game 3, Candidates match 1974 White to move Spassky has just played the surprising move 29...♘e7-f5!. It bears a strong resemblance to 16...♘c6!? as in the previous game, but the difference is that this time, putting psychology aside, it is the best move. As in the game against Averbakh his thought must have been that he cannot play normal positional moves because then he will surely suffer defeat because he is positionally lost. So the only thing he can do is to confuse Karpov and bring the game into a tactical domain where it’s easier to make mistakes in calculation. The question now is whether Karpov should accept the sacrifice or ignore it and pretend that nothing has happened? Karpov played 30 ♗xg5!. Karpov thought for only four minutes before he decided on not only the safest but also the best move. Gunnar Johansson mentions in his book Schackgeniet Anatolij Karpov that it’s possible White wins if the sacrifice is accepted: 30 exf5 e4 31 ♗d2 exf3+ 32 ♗xf3 gxf5. I have checked this with Komodo 11 and certainly it’s a win, for example after 33 ♗h5 ♖c8 34 b5 etc., but Karpov’s move is the most pragmatic and the most precise. 30...♘d4 Or 30...♗h6 31 exf5 ♗xg5 32 ♘e4. 31 bxc5 ♘xc5 32 ♖b6 ♗f6 33 ♖h1+! ♔g7 33...♔g8 is met by 34 ♖xd6 ♗xg5 35 ♖xg6+ ♔f7 36 ♖xg5. 34 ♗h6+ ♔g8 35 ♗xf8 ♖xf8 36 ♖xd6 ♔g7 37 ♗d1 ♗e7 38 ♖b6 ♗d8 39 ♖b1 ♖f7 40 ♘a4 ♘d3 41 ♘b6 g5 Black tries to flounder with ...g5-g4 but of course it doesn’t work against a giant like Karpov who plays with the highest precision. 42 ♘c8! ♘c5 42...g4 43 ♘d6 gxf3+ 44 ♔f1 with a double threat on f7 and b7. 43 ♘d6 ♖d7 44 ♘f5+ ♘xf5 45 exf5 e4 46 fxe4 ♘xe4 47 ♗a4 ♖e7 48 ♖he1 ♘c5 49 ♖xe7+ ♗xe7 50 ♗c2 ♗d8 51 ♖a1 ♔f6 52 d6 ♘d7 53 ♖h1 ♔e5 54 ♖d1 ♔f4 55 ♖e1! Black resigned. In accordance with his style Karpov chose to play the pragmatic ♗xg5 on move 30, which also happened to be the best move. Here one could also consider Nimzowitsch’s statement that if you can choose between a complicated continuation which might lead to an advantage and a safe positional continuation which will lead to an advantage, then you should always choose the latter. We will end our section about openings and middlegames by looking at one more recent example where pragmatism is adopted as an overall strategy. 150 Giri – Anand Stavanger 2017 White to move 12 a4!? Giri writes in his comments in New in Chess Magazine (5/2017) that he had several options and also considered 12 d4, 12 d3 and 12 h3. He thought that these three moves were too compromising and thus preferred the text move so he could see how Anand would react in reply. At the same time he could gain some time on the clock since most likely Anand wouldn’t have prepared for this move in his previous calculations. 12...b6 13 d3 ♗g4 14 h3 ♗h5 Anand is normally fond of exchanging his bishop for a knight but here he wants something more than just the knight and that is a weakening of his opponent’s kingside. 15 g4 ♗g6 16 ♘h4 16...♖c8 Giri was ready to meet 16...♘xg4 17 ♘xg6 ♘h6! with 18 ♗f3 ♖c8 19 ♘xe5 dxe5 20 ♗e4 followed by ♕f3 when it is White who is ruler of the game, since the h6-knight cannot reach h4 via f5. 17 ♘xg6 In his annotations to the game Giri admits that he didn’t want to sit and calculate the consequences of ...♘xg4 after every move and therefore quickly made this exchange. Pragmatic chess! 17...hxg6 18 ♗f3 g5 18...♘d7 would have made it more difficult for White to open the game for the bishop pair, e.g. 19 ♗c3 (or 19 ♗a3 ♘c5 20 ♗d5 ♘e7 21 ♗g2 ♘c6) 19...♘c5 20 ♗d5 e4!? or 20...♘e7. 19 ♗g2 ♘d7 20 f4!? gxf4 21 exf4 21...♘c5? Black’s best continuation was the concrete variation 21...♕h4! 22 fxe5 dxe5 23 ♔h2 ♘c5 24 ♗d5 ♘e6 25 ♕f3 ♘g5 26 ♗xf7+ ♔f8! 27 ♗a3+ ♖e7! with dynamic equilibrium. 27...♘e7? would be a mistake due to 28 ♗d5+!! ♘xf3+ 29 ♖xf3+ ♕f6 30 ♖af1. 22 fxe5 dxe5 23 ♗d5! Black had so many troubles that Anand had to resign after 33 moves. The bishop pair was simply too strong. What’s interesting here is how pragmatic Giri behaved in this phase of the game. Especially moves 12, 17 and 20 showed that it’s not necessarily all about the best move but rather the practically best move. Let’s move on to the second section of the book and concentrate on the final phase of the game, the endgame! Part 2: 150 most important positions in the Endgame 151-167: Pawn Endings Pawn endings are the foundation of all endings and a mastery of these is mandatory. One of the first things to learn is the opposition, especially the distant opposition. We will describe this idea with a few simple examples to be sure it is fully mastered in all its different aspects. Another important theme is triangulation which also touches on the subject of corresponding squares. The next theme is the square but in an advanced form. It also touches on the interesting area of illusions in chess. We will end the section on pawn endings by explaining how to play in positions with one extra pawn on the wing and a pawn majority. 151 Gligorić – Fischer Candidates Tournament, 1959 Black to move 57...♔b8! After this move, keeping the distant opposition, the players agreed to a draw. The white king cannot go to b4 and claim the opposition since the pawn stands in its way. 57...♔b7? obviously loses to 58 ♔b5! as well as 57...♔c7? when after 58 ♔c5! ♔b7 59 ♔b5! White gains the opposition. Black cannot prevent White from reaching c6 or a6, where it will escort the pawn to its promotion square. If White wants to test Black a little a likely continuation is 58 ♔d5 ...but after 58...♔b7! it’s still a draw. For example, following 59 ♔c5 ♔c7 60 ♔b5 ♔b7 61 ♔a5 ♔a7 62 ♔b5 ♔b7 63 ♔c5 ♔c7 Black continually retains the opposition so White is forced to advance the pawn. 64 b5 ♔b7 65 b6 ♔b8 66 ♔c6 ♔a8 66...♔c8 67 b7+ ♔b8 68 ♔b6 also leads to stalemate. 67 ♔c7 and Black is stalemated. 152 Mandler 1969 White to move The only way to hold the draw is to exploit the fact that the gpawn is placed on g6 which means that Black’s king cannot be placed there. 1 ♔b2! White will eventually trick Black with the distant opposition. 1...♔b6 1...♔b5 is answered by 2 ♔b3 immediately gaining the opposition. 2 ♔c2 ♔c6 3 ♔d2 ♔d6 4 ♔e2 ♔e6 5 ♔f2 ♔f6 6 ♔g2 This is the point. Black would like to place his king on g6 but it’s obstructed by its own pawn. 153 Grigoriev Shakhmatny Listok 1931 White to move The only way to hold the draw is by means of the opposition. When the black king captures the white pawn White’s king must be on b4. This is how it should be done: 1 ♔g3! ♔c2 Black wants to capture the pawn while shouldering the king. 2 ♔f2! If White attacks the b7-pawn he will lose after 2 ♔f4 ♔d3 3 ♔e5 ♔c4 4 ♔d6 ♔b5 5 ♔c7 ♔a6. 2...♔d2! 2...♔d3 3 ♔e1 ♔c4 4 ♔d2 ♔b5 5 ♔c3 demonstrates nice geometry with a king walk along the diagonals from h4-e1-b4 which is the only route to play for a draw. 3 ♔f1! ♔d3 4 ♔e1 ♔c4 5 ♔d2 Now White only follows the black king on the diagonal to b4. 5...♔b5 6 ♔c3 ♔xb6 6...♔c5 7 ♔b3! doesn’t help Black. 7 ♔b4 White has the opposition and the position is a draw. 154 Mattison 1918 White to move 1 g6! The only way to keep the draw is by sacrificing the pawn. 1 ♔h2?? fails to 1...♔g4 2 g6 fxg6 3 f5 ♔xf5! (3...gxf5 4 ♔g2 ♔f4 5 ♔f2 ♔e4 6 ♔e2 is a draw, since White has the opposition.) 4 ♔g3 ♔g5 and Black wins by obtaining the opposition. 1...fxg6 2 f5! It’s surprising that White must sacrifice both pawns to achieve the draw. 2...gxf5 3 ♔g1! White holds the draw by distant opposition. 3 ♔h2?? would give Black the opposition by 3...♔h4! 4 ♔g2 ♔g4 5 ♔f2 ♔f4 6 ♔e2 ♔g3 and Black’s pawn is escorted to promotion. 3...♔g5 4 ♔f1 If the black pawn had been on f6 Black would have gained the opposition by placing the king on f5. Now that isn’t possible and White retains the opposition. 4...♔f4 5 ♔f2 ♔e4 6 ♔e2 f4 7 ♔f2 f3 8 ♔f1 ♔e3 9 ♔e1 f2+ 10 ♔f1 ♔f3 It’s stalemate. 155 Fahrni – Alapin Das Endspiel im Schach (Leipzig 1917) White to move This so-called Fahrni endgame probably took place in Munich some time between 1909-1917 but there is no evidence of this. The only thing we know for sure is that Fahrni published this ending on page 15 in his endgame book. This position has arisen in many games, for example Fiebig – Tartakower, Barmen 1905, where Tartakower was lucky because Fiebig failed to win. The trick to win is to reach this position with Black to move. The only way to achieve this is by triangulation. You can play the king either to the squares d5-d4-c4 or d5-c4-d4 and in both cases the three squares are in the shape of a triangle. If Black’s king is placed on c7 as in the diagram, White wants to place the king on the corresponding square c5 and then break through to b6 next move and pick up the a6-pawn. The main theme of this position is not only to show the trick how to lose a tempo but also to introduce a simpler form of corresponding squares. The d5-square corresponds to the c8-square and the c4-square to the c8-square and that means that the player who is to move is at a disadvantage. 1 ♔d5 ♔c8 2 ♔d4 Another way to triangulate is 2 ♔c4 ♔d8 3 ♔d4. 2...♔d8 Or 2...♔b8 2 ♔c4 leading to the same thing. 3 ♔c4 ♔c8 4 ♔d5! So it’s Black’s turn to move. He loses after the further: 4...♔d8 5 ♔d6 ♔c8 6 c7 ♔b7 7 ♔d7 ♔a7 8 ♔c6! The strongest move. Beware, 7 c8♕?? at once is stalemate! 8...♔a8 9 c8♕+ ♔a7 10 ♕b7 mate So don’t forget this important triangulation manoeuvre. The young Kasparov didn’t miss it when he played Alburt in Daugavpils, 1978, but Alburt was “impolite” and resigned one move too early, not allowing Kasparov to complete the triangulation. Triangulation is a more common device than one might believe. The idea of losing a move and placing the opponent in a zugzwang situation can be found in other endings as well, for example in the ending queen against rook, position 284. The concept of corresponding squares is of immense importance and we’ll touch on it again in the two following positions, as well as in the section Bishop Endings, position 189. 156 Moravec 1940 White to move Black’s d5-pawn is doomed but White must capture it in the right way. Correct is 1 ♔b4! 1 ♔c3? ♔g5 2 ♔d4 ♔f4 leads to a draw as White is forced to make a move. Compare this with the main variation where the same position arises but with Black to move. 1...♔g5 2 ♔c5 ♔f4 3 ♔d4! The d4-square corresponds to the f4-square and White wins due to zugzwang. Of course not 3 ♔xd5? ♔e3 and Black captures the pawn. After the further 3...♔g3 4 ♔xd5 ♔f2 5 e4 the e-pawn moves two squares in one move. Note that the position is a draw if we move all the pieces up one square. The reason is that White can then only move the pawn one square since it stands on e3 while the black king on f3 is ready to capture it. 157 Lasker 1900 White to move This position is a more advanced form of corresponding squares. White can break through via c4 to b5 or via h4 to g5 or h5. The key position is when the king is on d3. For White to win, the black king must be on c7 with White to move. The solution is 1 ♔b2!. 1 ♔b3? is a draw after 1...♔a7!. 1...♔a7 1...♔b8 is answered by 2 ♔c2! ♔c8 3 ♔d2! ♔d8 4 ♔c3! ♔c7 5 ♔d3! and Black is in zugzwang. 2 ♔b3! The b3-square corresponds to the a7-square. 2...♔b7 Or 2...♔a6 3 ♔c2!. 3 ♔c3! ♔c7 4 ♔d3! ♔d7 Or 4...♔b6 5 ♔e3 ♔c7 6 ♔f3 ♔d7 7 ♔g3 ♔e7 8 ♔h4 ♔f6 9 ♔h5. 5 ♔c4 and White wins. It’s important to grasp this study so go through it carefully until you understand it 100 percent. If you understand this position you will understand other positions where corresponding squares are involved. Why not look at position 197 which also touches on the same theme? 158 Réti 1921 White to move Why is this pawn ending important? You will never reach this position in a game but it’s the ideas it represents that are important. The position is a very instructive and pure example showing how to play with multiple plans or double threats with every move. White has two goals. He wants either to support his pawn or catch the enemy pawn. On the surface it looks impossible but the geometry of chess can sometimes fool you. 1 ♔g7! By placing the king on this square White is following two paths at the same time and accordingly has two plans in mind. 1...h4 1...♔b6 2 ♔f6 ♔xc6 3 ♔g5 and the black pawn is lost. 2 ♔f6 ♔b6 2...h3 doesn’t win on account of 3 ♔e6 or 3 ♔e7, escorting the passed pawn to c8. 3 ♔e5!! The double threat of 4 ♔d6, supporting the c6-pawn, and 4 ♔f4, trapping the enemy pawn, secures half a point. The endgame expert Averbakh used this position as an example in Chess Tactics for Advanced Chess Players to explain that the double threat is the heart of all tactics and combinations. 159 Moravec Československý Šach 1952 White to move This position is illusory since the white king is outside the square of the enemy pawn. Nevertheless it’s still a draw. White approaches both pawns by playing 1 ♔b5. Black replies 1...h5. Of course if 1...♔c7 White gains the necessary tempo to reach the square h5-h1-d1-d5 after 2 ♔c4 h5 3 ♔d5. 2 ♔c6!! White sets up a double threat and achieves the draw. This move mimics 3 ♔e5!! in the previous study by Réti but the beautiful point is that in Moravec’s study the king is placed in front of the pawn, setting up the necessary double threat of ♔b7 and ♔d5 to achieve a draw. 2...♔c8 After 2...h4 3 ♔b7 h3 4 c6 h2 5 c7+ ♔e7! both pawns promote. 3 ♔d5 White is within the square h5-h1-d1-d5 and catches the pawn just in time before ...h1♕ after the further 3...h4 4 ♔e4. The square gets smaller and smaller. 4...h3 5 ♔f3 h2 6 ♔g2 h1♕+ 7 ♔xh1 160 Réti 1928 White to move It’s unbelievably magical that this position is a draw, despite the fact that White is two pawns down, although thanks to position 158 it’s easier to understand that the miracle of Dunkirk is indeed achievable. 1 ♔g6 ♔b6 1...h5 2 ♔xg7 h4 3 ♔xf6 leads to exactly the same position as the previous example by Réti. 1...f5 2 ♔xg7 f4 3 ♔f6 f3 (3...♔b6 4 ♔e5) 4 ♔e6 or 4 ♔e7 is a draw as well. 2 ♔xg7 h5 2...f5 3 ♔f6 f4 4 ♔e5 f3 5 ♔d6 is likewise a draw. 3 ♔xf6 h4 4 ♔e5 and again we have the same position as in Réti’s magnum opus. Don’t forget illusions and miracles in chess! The great chess teacher GM Luděk Pachman once said that there are no miracles in chess and to a certain extent he’s right, but nevertheless we have to be prepared for the extraordinary when it occurs, whether we solve studies or play a game. It’s sometimes possible to run through a wall and we have to understand when that is indeed a reality – since magic sometimes interweaves with real life. 161 Schlechter – Marco 1893 White to move On the surface the a-pawn looks doomed whereas the black dpawn will eventually queen. However, this is a famous illusion which has been seen many times in practical games and studies. It’s said that the famous study by Réti from 1921, position 158, was inspired by this position which appeared in 1893 in Deutsches Wochenschach. The game was agreed drawn after 57 a4 ♔b4 58 ♔b6! since the white king will support the pawn if a4-a5 follows, while 58...♔xa4 is answered by 59 ♔c5. This is the idea in its simplest form and the key idea to remember when dealing with more advanced examples. 162 Rinck 1922 White to move This looks to be an easy win, but after 1 a4 ♔b3 2 a5 ♔c3! White must be careful. 2...♔c4 seems to lead to the same situation we saw in Schlechter – Marco, where Black is threatening to enter the square of the apawn while threatening to support his own pawn. However, White wins after 3 a6 ♔d3 4 a7! f2 5 a8♕ f1♕ thanks to the skewer 6 ♕a6+. 3 ♔g1! White answers Black’s clever move with an equally clever move. 3 a6 doesn’t work on account of 3...♔d2! 4 a7 f2 5 ♔g2 ♔e2 when both pawns promote. Notice that 3 ♔g3? ♔d4! would have been a copy of Schlechter – Marco, or if you prefer the famous Réti study where ♔e5 was the key move. After 4 a6 ♔e3 5 a7 f2 it’s a draw. 3...♔d4 4 a6 ♔e3 5 ♔f1! This is the point. White’s pawn promotes while Black’s is blockaded. 163 Lasker – Tarrasch Saint Petersburg 1914 White to move In the famous tournament book of Saint Petersburg 1914 Tarrasch commented that White seems to be lost. 40 h4 ♔g4 41 ♔g6! The saving problem move which echoes the move Schlechter played against Marco when he went 58 ♔b6!. The natural 41 ♔f6? loses by force after 41...c4 42 bxc4 bxc4 43 ♔e5 c3 44 bxc3 a4 45 ♔d4 a3 since White’s c3-pawn blocks the path for the king to stop the black runner. 41...♔xh4 42 ♔f5 The correct path is to walk along the b1-h7 diagonal. 42...♔g3 42...c4 43 bxc4 bxc4 44 ♔e4 c3 45 bxc3 ♔g5 (45...a4? loses to 46 ♔d3 a3 47 ♔c2 and White catches the pawn while supporting his own.) 43 ♔e4 ♔f2 44 ♔d5 ♔e3 45 ♔xc5 ♔d3 46 ♔xb5 ♔c2 47 ♔xa5 ♔xb3 and the game was agreed drawn. 164 Lolli 1763 Black to move The golden principle to hold the draw a pawn down in this type of position is to be as active as possible with the king and then advance the h-pawn to secure a profitable exchange. 1...♔f4 The principled move, although other king moves on the f- and gfile hold as well. The main thing for Black is to prevent the opponent’s king from reaching h6 and in this position this goal will be achieved if the white king reaches h4 or g4. 1...♔e4? is a mistake due to 2 ♔g3 ♔f5 3 ♔h4. If White’s king reaches h4 or g4 it’s a win due to the availablity of tempo-gaining moves with the pawns. For example 3...♔g6 4 ♔g4 h6 (Or 4...♔f6 5 ♔h5 ♔g7 6 ♔g5 ♔h8 7 ♔h6 165 Black to move This is a well-known winning position where White wins by advancing the pawns and playing his g-pawn to g6 at the right moment. 7...♔g8 8 g4 ♔h8 9 g5 ♔g8 10 h3! The idea is to save a tempo and play the pawn to h4 only when Black’s king is on h8. 10...♔h8 11 h4 ♔g8 12 h5 ♔h8 13 g6 hxg6 14 hxg6 ♔g8 15 g7 ♔f7 16 ♔h7 and White queens.) 5 ♔f4 ♔f6 6 ♔e4 ♔e6 7 h4 ♔f6 8 g3! (8 g4? is a draw after 8...♔e6.) 8...♔e6 9 g4 (With Black to move it is a loss since White penetrates with the king next move.) 9...♔f6 10 ♔d5 ♔e7 11 ♔e5 ♔f7 12 ♔f5 ♔g7 13 ♔e6 ♔g6 14 h5+! ♔g5 15 ♔f7 ♔xg4 16 ♔g6 and White wins the h6-pawn and the game. 2 g3+ 2 h3 is answered by 2...h6! (2...h5? weakens the g5-square as can be seen after 3 g3+ ♔e4 4 ♔e2 ♔d4 5 ♔f3 ♔e5 6 ♔e3 ♔d5 7 ♔f4 ♔e6 8 ♔g5.) 3 g3+ ♔f5 (After the more active 3...♔e4 4 ♔e2 Black must move his king to the fifth rank anyway.) 4 ♔f3 h5!. The only move to avoid a penetration by the white king. Now Black easily holds with the help of the opposition: 5 ♔e3 ♔e5 6 ♔d3 ♔d5 and White cannot improve his position any further. 2...♔g4 The simplest, although 2...♔e4 is also a draw since the king cannot go to g3. 3 ♔g2 h5 Of course 3...h6 is a draw too. 4 h3+ 4...♔g5! When the black pawn is on h5 this is the only move, since White will penetrate after 4...♔f5? 5 ♔f3 ♔e5 6 ♔e3. 5 ♔f3 ♔f5 and White cannot win. The golden rule to remember for the defender is not to allow the attacker to reach the fourth rank. 166 Euwe Bondeslutspel 1946 White to move The pawn structure is the same that arises in the exchange variation of the Spanish Game after 1 e4 e5 2 ♘f3 ♘c6 3 ♗b5 a6 4 ♗xc6 dxc6. This structure also arises in the Caro-Kann after 1 e4 c6 2 d4 d5 3 ♘d2 (or 3 ♘c3) 3...dxe4 4 ♘xe4 ♘f6 5 ♘xf6+ exf6 but on the kingside. In both variations Black must be careful not to allow the exchange of all the pieces as this will lead to a lost pawn ending. The excellent variations by Euwe show how to win most convincingly. White has four healthy pawns versus Black’s three on the kingside, while Black has four against three on the queenside. However, the problem for Black is that his c-pawns are doubled and with correct defensive play by White on the queenside it’s impossible to create a passed pawn. Let’s see how this strategy is outlined by Euwe. 1 ♔d2 The first step for both players is to activate their kings. 1...♔e7 2 ♔e3 ♔e6 3 f4 The second step involves a combination of activating the pawn majority as well as halting the opponent’s majority as early as he can. 3...c5 4 c4! White stops any further advance of the c-pawn at once, thereby preventing Black from gaining time and space. 4...c6 5 a4 White threatens the crippling a4-a5. 5...b5 6 b3! Of course White keeps his pawns intact on the queenside. 6...f6 7 a5 b4 8 g4 8...g5 Euwe doesn’t mention the passive defence 8...♔d6. White can play according to the same formula as in the main variation by Euwe but there is one pitfall to be aware of and that’s 9 e5+! fxe5 10 f5. Now it looks like the pawn majority will soon be transformed to an outside passer which will decide the game. This is true but after 10...♔e7 11 g5 ♔d7 12 ♔e4 ♔d6 one has to be careful. Here it’s important to use the tempi available in the h-pawn with 13 h3! (13 h4?? g6 is a draw. It isn’t possible to triangulate with the king after 14 f6 ♔e6 15 ♔f3 ♔d6 16 ♔e3 ♔d7! because Black has a triangle as well on the e6-d6-d7 squares.) 13...g6 14 h4 and Black is in zugzwang and loses the e5-pawn. Back to the main variation by Euwe: 9 e5! The e4-pawn, the child of the sun, has to be sacrificed to secure a pawn majority on the kingside. 9...gxf4+ After 9...fxe5 10 fxg5 White wins more quickly. 10 ♔xf4 fxe5+ 11 ♔e4 White has achieved his dream position and it shows the advantage of a distant pawn majority and eventually the distant passed pawn in its clearest light. 11...h6 12 h4 ♔f6 13 g5+ hxg5 14 hxg5+ ♔xg5 15 ♔xe5 ♔g4 16 ♔d6 ♔f4 17 ♔xc6 ♔e4 18 ♔xc5 ♔d3 19 ♔xb4 ♔d4 The simplest way to win this position is by triangulation. White can triangulate on the squares a3-a4-b4 or a4-a3-b4. 20 ♔a3 The alternative triangulation goes 20 ♔a4 ♔c5 21 ♔a3 ♔d4 22 ♔b4 winning, but not 20 c5?? ♔d5 with a draw. 20...♔c5 21 ♔a4 ♔d4 Black must prevent the pawn move to b4. 22 ♔b4 ♔e5 23 ♔c5 and White promotes one of the pawns. An instructive ending worth playing over with a friend or a computer a couple of times from both sides of the chesboard. You will then discover other interesting variations. Before we end this section on pawn endings and move on to knight endings we must not forget the most incredible pawn ending I’ve ever seen. I discovered it with delight many years ago when I studied pawn endings in Reuben Fine’s Basic Chess Endings. 167 Berger – Bauer Correspondence 1889-1891 White to move Sometimes it’s possible to win even though the pawn majority is crippled. This instructive position arose in a correspondence game a long time ago. If White wants to win he must participate in a hair raising pawn race since that’s the only way to win. Exact calculation is required and in over the board play it wouldn’t be so easy to find the ingenious win achieved by Berger. He played 1 c4!! bxc3 2 ♔e3 ♔g5 So what is going on? Here comes the next shocker! 3 a4!! 3 ♔d3 is a draw after 3...♔xg4 4 a4 h5 5 b4 h4 5 b5! and both pawns promote. However, after shocker number two it’s a different story because Black is soon forced to capture the b-pawn and that makes a difference. White gains a crucial tempo in the pawn race. 3...♔xg4 4 b4 axb4 5 ♔d3!! Another surprise move. How many are there in this position? 5 a5 suprisingly leads to a draw after 5...b3 6 ♔d3 b2 7 ♔c2 ♔f3!. This is the point, Black concentrates on promoting his c-pawn. 8 a6 ♔e2 9 a7 b1♕+ 10 ♔xb1 ♔d2 11 a8♕ c2+ 12 ♔a2 c1♕ 13 ♕g2+! The queen ending is an easy draw. 5...h5 6 a5 h4 7 a6 h3 8 a7 h2 9 a8♕ and White wins handsomely by following up with the deadly ♕h1. The advantage with promoting a rook pawn is that it automatically controls the opponent’s corner on the long diagonal so never underestimate the cheapest material on the board, since it can suddenly turn out to be the most valuable piece on the board! 168-180: Knight Endings Knight endings have some similarities to pawn endings which is useful to keep in mind whenever we play or study a knight endgame. The same type of advantage that often proves decisive in pawn endings is often decisive in knight endings, whether it’s an outside passed pawn or the exploitation of four pawns against three on the same side. 168 Grigoriev 1932 White to move One of the worst enemies for the knight is the rook pawn (and the second worst is the knight pawn) but in some positions the knight can turn out to be a magical piece indeed. In the study by the great study composer Grigoriev White saves the game after: 1 ♘f7! h3 2 ♘g5 h2 3 ♘e4+ 3...♔c2 3...♔d3 4 ♘g3! leads to an important position to remember. White’s knight acts like a wall preventing Black’s king from approaching the knight by 4...♔e3 or 4...♔d2 due to the fork 5 ♘f1+. 3...♔d4 4 ♘f2! leads to another impregnable wall since 4...♔e3 fails to the fork 5 ♘g4+. 4 ♘g3! 4 ♘f2? is the wrong route to take as after 4...♔d2 5 ♔d6 ♔e2 6 ♘h1 ♔f3 7 ♔e5 ♔g2 8 ♔f4 ♔xh1 White is one tempo short reaching f2, imprisoning the king in the corner. 4...♔d1 5 ♔d6 ♔e1 6 ♔e5 ♔f2 7 ♔f4 ♔g2 8 ♔g4 and White draws. The knight’s possibilities of extending the wall by means of forks is the important idea to remember. 169 Jaenisch 1837 White or Black to move Normally the possessor of one or several pawns can play for a win unless there is a rook pawn involved and the king is in front of it. In such a situation it is worthwhile for the king and knight to find out if it’s possible to organise the pieces in such a way as to checkmate the king in the corner. In this study by Jaenisch White wins whether it’s his turn to move or not. This is how to do it: First the king goes to f1 while Black advances his pawns. 1 ♔g3 f5 2 ♔f2 h5 3 ♔f1 f4 4 ♔f2 h4 5 ♔f1 h3 The moment has come to manoeuvre the knight. 6 ♘e5! ♔h2 7 ♔f2 ♔h1 8 ♘g4 White forces another pawn move with the help of the knight. 8...f3 9 ♔f1 f2 10 ♘xf2+ ♔h2 11 ♘e4 ♔h1 12 ♔f2 ♔h2 13 ♘d2! This regrouping of the knight is the important idea to remember. 13...♔h1 14 ♘f1! h2 15 ♘g3 mate. 170 Laroche – Greville Paris 1848 White to move This endgame was played at the famous Café de la Régence in its heyday before it moved to Hôtel Dodun in 1852. Kieseritzky remarked that the ending was very interesting. I can only agree since I’m very well acquainted with this endgame after having played several training games from exactly the same starting position, more than 20 years ago. The trick to hold the draw is fairly simple. Black should place his king in the middle area of White’s pawns, that is the g6-square, and when one of the pawns is pushed to the fifth rank Black will blockade it with the king. It’s of immense importance not to allow all the pawns to reach the fifth rank. A knight is worth three pawns only if these are located on the fourth rank. Placed on the fifth rank they outsmart the knight. 1 ♔g4 ♔f6 2 h4 My opponent during the training session, Morgan Lundqvist, played 2 ♔h5 preventing Black from reaching g6 with his king. However, this is of no consequence because after 2...♔f5 the white king is too misplaced to help in supporting the pawns. 2...♔g6 3 ♔f4 ♘d5+ The simplest way to draw was 3...♔h5! and then hit the weakest link in the pawn chain with the knight. Plausible lines are 4 ♔e5 (4 f3 ♘d5+ 5 ♔e4 ♘c3+ 6 ♔f5 ♘e2) 4...♘c6+ 5 ♔d5 ♘b4+ 6 ♔e6 ♘c6 7 f4 ♘d4+ 8 ♔e5 ♘e2 with a draw. 4 ♔e5 ♘b4 Black plans to go behind the pawns. 5 f3 ♘d3+ 5...♔h5! was still the simplest. 6 ♔e4 ♘c5+ 7 ♔d4 ♘e6+ 8 ♔e5 ♘f8 9 g4 Black has played unnecessarily passively with the king and the knight but amazingly the position is still a draw with precise play. 9...♘d7+? This is the losing move. Correct was 9...♔h6!! followed by ...♘g6+ 10 h5 is answered by 10...♔g5. 10 ♔f4? White had a forced win with 10 ♔d6 ♘f8 (10...♘f6 11 ♔e6) 11 ♔e7 ♔g7 12 h5 ♘h7 13 f4 ♘f6 14 h6+ ♔g6 15 f5+ ♔g5 16 ♔f7. 10...♘c5 11 ♔e3 ♘e6 12 f4 ♘g7? Yet another losing move. It was a draw after the more active 12...♘c5 13 ♔d4 (13 ♔f3 ♘d3 14 ♔g3 ♘e1) 13...♘d7! preventing the king from penetrating the centre. Then 14 ♔e4 ♘f6+ 15 ♔f3... 171 White and Black to move Precisely this position has been elaborated in detail by Averbakh/ Chekhover in their book Knight Endings (Batsford 1977). Black to move draws after 15...♘d5 16 f5+ (or 16 h5+ ♔h6 17 ♔e4 ♘f6+ 18 ♔f5 ♘d5 19 ♔e5 ♘e3) 16...♔f6. The main principle is to place the king in front of the checking pawn. 17 ♔e4 ♘c3+ 18 ♔e3 (18 ♔d4 ♘e2+ 19 ♔e3 ♘g3 20 ♔f4 ♘e2+) 18...♔e5!!. The only move to draw. Notice the active play by the black king. 19 h5 ♘d5+ 20 ♔f3 ♘f6 21 h6 ♘h7 Black’s pawns are blockaded and a draw is inevitable. White to move wins in the diagram above after 1 f5+!. The rule given by Averbakh/Chekhover states that “three connected pawns defeat a knight if they all reach the fifth rank.” But, as we’ll discuss later, it’s important that the pawns are well supported by the king because otherwise the knight can destroy the nice pawn phalanx on the fifth rank by threatening one of the pawns that cannot be protected by the king. Play continues 1...♔g7 2 g5 ♘d5 3 h5. All three pawns have reached the fifth rank and Black cannot prevent their further advance. 3...♘c3 (If White manages to set up a pawn chain after 3...♔f7 4 h6 ♘c3 5 h7 ♔g7 6 g6 it’s impossible to prevent the decisive f5-f6+ at the right moment.) 4 ♔f4 ♘e2+ 5 ♔e5 ♘g3 6 f6+ ♔g8 7 h6 ♘h5 8 g6. The pawns have reached the sixth rank. After 8...♘g3 9 h7+ ♔h8 10 f7 one of the pawns is promoted. 13 ♔e4 ♘e8 14 f5+ ♔f7 15 g5 ♘d6+ 16 ♔f4 ♘e8 17 h5 The pawn roller crushes Black. 17...♘g7 18 h6 ♘e8 19 ♔e5 ♔g8 20 g6 ♔f8 21 f6 ♘c7 22 h7 Black resigned. 172 Hillarp Persson – Dragnev Stockholm 2018 White to move In this position the Swedish grandmaster Tiger Hillarp Persson played what was to my eyes the amazing move 54 ♔g5!!. I was surprised because I was sure he was going to play 54 f5, which is actually a mistake. Let me explain. I grew up with the endgame books by Fine and Averbakh so when I saw this position during a visit to the Rilton Cup, on the first day of 2018, I was staggered Tiger didn’t play 54 f5 since, according to Fine’s rule in Basic Chess Endings, this endgame is won if two pawns reach the fifth rank. Later in the 70s this rule was altered by Averbakh and Chekhover. Their new rule stated that all pawns must be placed on the fifth rank to secure a win. If we apply this rule in the present position it’s actually a draw. Let’s try 54 f5?. Black activates the knight either with 54...♘f2! or 54...♘b2! and holds the draw even when all pawns are placed on the fifth rank after 55 g5. What is going on? Black attacks the e5-pawn with 55...♘d3! and after 56 e6 ♘f4 (There are three other good knight moves too.) 57 ♔g4 ♘e2! we have the following position: 173 White to move Obviously White wants the king on e5 but that is not possible after 58 ♔f3 (58 ♔h5 is met by 58...♘d4! 59 ♔g6 ♘f3! when it’s also a draw due to the alternating threats against g5 and f5.) due to 58...♘d4+ 59 ♔e4 ♘xe6. This is the most illuminating variation showing why the position is a draw. The reason White was unable to win, despite the fact that all the pawns were located on the fifth rank, was that Black could force one of the pawns to the sixth rank and in this case it secured the draw. Averbakh/Chekhover were not completely correct in stipulating that “three connected pawns defeat a knight if they all reach the fifth rank” since it’s vital that the defender cannot disrupt the horizontal formation of the pawns on the fifth rank, which actually is the case here. The necessary and simplest supplement to Averbakh’s/Chekhov’s rule is to reformulate it by stating that “three connected pawns on the fifth rank win against a knight if they are supported by the king.” 54...♔f7 55 f5 Now it’s easier to prepare the push of the g-pawn to the fifth rank since the white king is close to all the pawns. 55...♘f2 56 e6+ Note that Fine’s rule with two pawns on the fifth rank doesn’t work even here, despite the fact that these pawns are fully supported by the king! After 56 ♔f4 follows 56...♘d3+ 57 ♔e4 ♘f2+ 58 ♔f3 ♘h3 and White is unable to improve his position since the g-pawn cannot reach the fifth rank. All winning tries leads to a blockade one way or the other. 56...♔e7 57 ♔h5! As we have seen before 57 ♔f4? ♔f6 is no use, since it doesn’t help White to play 58 g5+ because after 58...♔e7 it’s a draw. 57...♘d3 58 g5 ♘e5 59 ♔h6 ♘g4+ 60 ♔g7 ♘e3 61 ♔g6 White’s triangulation forces Black to face a zugzwang. 61...♘c4 62 ♔h7 ♘e3 63 f6+! ♔xe6 64 g6 ♘g4 64...♔xf6 loses to 65 g7. 65 f7 ♘f6+ 66 ♔g7! ♘h5+ 67 ♔h6 Black resigned. Well played by Hillarp Persson who consciously or unconsciously played contrary to the general rules formulated by the endgame giants Fine, Averbakh and Chekhover by using exact calculation. He has left an important contribution in how this endgame should be played and understood. 174 Halberstadt 1952 White to move Superficially this position looks like a dead draw. After all, Black’s knight on f8 stands in front of the passed pawn and the king is protecting the knight. So, we have another illusion. Chess is full of them and we must learn them one by one! 1 ♘f4! Such a knight manoeuvre, when the attacker dominates and controls the opponent’s knight, is an important idea to remember whether it’s an opening, middlegame or endgame. 1...♘h7 1...♔g7 allows the decisive exchange 2 ♘e6+. 2 ♘e6! Note how this knight persists in dominating its black counterpart, but now also the black king is under the knight’s influence. 2...♔h8 3 f3! As we have seen in some pawn endgames, it can also prove to be useful in knight endings to have tempo moves in reserve. 3 f4? would actually lead to a draw after 3...♔g8 4 f5 ♔h8 5 ♔f7 (5 ♘f8 liberates the knight after 5...♘g5.) 5...♘g5+! 6 ♘xg5. Black is stalemate. 3 ♔f7? is met by 3...♘g5+ 4 ♔e7 ♘f3 when the blockade is too much to handle for White. 3...♔g8 4 f4 ♔h8 5 ♔f7 5 ♘f8, forcing an exchange, also wins but takes longer. 5...♘g5+ 6 fxg5 and Black is mated in five moves. Who could imagine that Black would be mated after 11 moves playing the best moves available? If this isn’t a miracle then what is? 175 Keres – Reshevsky Leningrad/Moskow 1939 White to move This position looks easier to win than it really is. Black is more active than White, who must find a second weakness to attack while avoiding too many pawn exchanges. 34 ♔f1 ♔e7 35 ♔e2 ♔d6 36 ♘c2 ♔e5 37 ♘e3 ♘b2 38 ♘d1 ♘a4 38...♘c4 39 ♔d3 ♘d6 40 ♘e3, threatening ♘c4+, gives White a flexible position which can easily be improved. 39 ♔d3 ♔d5 40 ♘e3+ ♔c5 41 ♘f5! White is hitting a second weakness, which is obviously the kingside, but the weakness needs to be firmly established. 41...g6 42 ♘h6 f5 43 ♘f7 ♔d5 44 ♘g5 ♘c5+ 45 ♔e3 h6 46 ♘f3? The second weakness is the h6-pawn and White should have attacked it by 46 ♘f7! ♔e6 (46...♘e4 47 ♘xh6 ♘xc3 48 ♔f4 ♘e4 49 f3 is to no avail.) 47 ♘xh6 ♔f6 48 ♔d4 and White wins. 46...g5 47 g3 White must prevent ...f5-f4 which gives Black too much space. 47...♘e4 48 ♘d4 ♘xc3? Black had a forced draw after 48...f4+ 49 gxf4 ♘xc3 50 ♘f5 ♔e6 51 ♘xh6 ♘d5+. 49 ♘xf5 h5 50 ♘g7 Fine has suggested 50 f4 g4 51 ♘g7 but Black has the strong defence 51...♘b5!, controlling the important flight squares d6 and c7 for the white knight, as can be seen after 52 ♘xh5 ♔e6 53 ♘g7+ ♔f6 54 ♘e8+ ♔f7. The knight is trapped and although White has three pawns it’s not enough after 55 f5 ♔xe8 56 ♔f4 ♔f7 57 ♔xg4 ♔f6 58 ♔f4 ♘d4 59 g4 ♘e2+ with a draw. 50...h4 51 gxh4 gxh4 52 f4 h3! 53 ♘f5 ♔e6 54 ♘g3 ♔d5 55 ♔f3 ♔d4 56 ♘h1 ♘d1 57 ♘f2 ♘e3 58 ♔g3 ♔d5 59 ♔xh3? White wins with 59 ♔h4! ♔e6 60 ♔g5. 59...♔e6 Now it’s a forced draw because the f4-pawn cannot be held. 60 ♔g3 ♔f5 61 ♔f3 ♘f1 62 h3 ♘d2+ 63 ♔e2 ♘c4 64 ♔d3 64 ♘d3 ♔e4 gives Black a very active position and there isn’t much White can do. 64...♘b6 and a draw was agreed. If Keres failed to win with a passed pawn then the next ending should be even more difficult to win. There is only one way to acquire the necessary confidence when playing such endings and that is by testing them against a friend or a computer, both as an attacker and defender. Only then will you gain a good understanding of knight endings. 176 Gebhardt – Bellman Levelezés 1996 Black to move Black has good practical chances of exploiting the extra pawn in a situation where there are four pawns versus three on the same side of the board. However, it’s not easy to defeat a computer. In his famous Basic Chess Endings (the main endgame book before Averbakh, Dvoretsky, Nunn, Müller and Tablebase entered the arena) Fine believes that this endgame is probably won. Dvoretsky thinks Fine is right as he wrote in Dvoretsky’s Endgame Manual. The defender, which in this case is White, has three main defences: a) The first defence is to play passively and wait while Black activates his pieces and pawns. This plan is not to be recommended as it’s too risky. b) The second defence is to play actively with the f-pawn. If Black already has his pawn on e5 White can try to exchange it by preparing f2-f4. c) The third defensive idea is to mobilise the g-pawn. It seems best to opt for the second defence and go on the offensive with the f-pawn and this is also the plan recommended by Fine/Dvoretsky. In the game Black played 31...♘d4 which for the moment prevents White from centralising his king. If 31...e5 32 ♔f3 ♔f7 33 ♔e3 White is ready for f2-f4 next move. 32 f4! White has completed the most effective arrangement of his kingside pawns by creating a wave. When playing a pawn down, four versus three on the same side, this is the most resistent set-up for the defender. Compare this with a rook ending where the most desirable formation is an opposite wave consisting of pawns located on f2, g3 and h4. Such waves are easy to remember if you know of their existence and above all understand their effectiveness! 32...♔f7 33 ♘d3 33...g5? Normally Black wants to avoid pawn exchanges because then the drawing chances would increase dramatically for White. The main purpose of the move played is to obtain a passer on the e-file if White exchanges pawns on g5. Note that the same result will arise if Black prepares the ...e6-e5 push. However, the problem is that when Black’s pieces are not sufficiently active it will be easy for White to set up a blockade and stop the pawn from reaching e4. The normal continuation would be the activation of the kings. After 33...♔e7 34 ♔f2 ♔d6 35 ♔e3 ♘c6 36 ♔e4 f5+ 37 ♔e3 ♔d5 the kings are optimally placed. Black tries to drive White further down the board with manoeuvres such as ...♘a5-c4+ and ...♔d4. The best way to learn to conduct this ending is to play it against a friend or a computer. I have done it myself many times since it’s one of my favourite endings to practice and study. 34 fxg5 fxg5 35 ♔f2 ♔f6 36 ♔e3 ♘f5+ 37 ♔f3 ♘d6 38 ♘c5 An excellent square for the knight, keeping control of the important e4-square. 38...h6 39 ♔g4 e5 40 ♔h5?! 40 ♔f3 is a simpler draw, trying to blockade the e-pawn for as long as possible. 40...e4! 41 ♔g4? 41 ♔xh6? also loses after 41...e3 42 ♔h5 e2 43 ♘d3 ♘c4! 44 ♘e1 (44 ♔g4 is effectively prevented due to 44...♘e5+.) 44...♔f5 45 h4 g4 46 ♔h6 ♔f6 47 ♔h5 ♘e3 48 ♔h6 ♘f1 49 ♔h5 ♔f5 and the g3-pawn is lost. The last chance to hold the draw is by exchanging as many pawns as possible, commencing with 41 h4!. Black can try 41...♔f5 and after the long and forced variation 42 hxg5 hxg5 43 g4+ ♔f6 44 ♘d7+! ♔e6 45 ♘f8+ ♔e5 46 ♔xg5 e3 47 ♘g6+ ♔e4 48 ♘h4 e2 49 ♘g2 ♔f3 50 ♘e1+ ♔f2 51 ♘d3+ ♔f1 52 ♔h5 ♘b7 53 g5 ♘c5 54 g6 ♘xd3 55 g7 ♘f4+ 56 ♔h6! e1♕ 57 g8♕ White is a piece down but nevertheless it’s still a theoretical draw. 41...♔e5 42 h4 ♔d4 43 ♘e6+ ♔d5 44 ♘c7+ 44 ♘f8 ♘f7!. 44...♔c4 45 hxg5 hxg5 46 ♘a6 ♘b5 47 ♔h3 e3 48 ♔g2 ♔c3 49 ♘c5 e2 50 ♘e4+ ♔c2 White resigned. The main rule to keep in mind for the defender is to conduct an active defence (32 f4!) and exchange as many pawns as possible (41 h4!). An additional resource for the attacker, because of his extra pawn, is that in some positions it’s possible to sacrifice the knight for two pawns. This means that the attacker plays with three pawns versus a knight which in some cases, as we already have mentioned, means a win. This is an idea to bear in mind for both the attacker and the defender! Good luck whenever you reach this fascinating but unfortunately rare ending. 177 Lasker – Nimzowitsch Zürich 1934 Black to move This position illustrates the advantage of having a distant passed pawn. It’s also a typical example for understanding the meaning of Botvinnik’s statement that knight endings are like pawn endings. However, in knight endings there are more defensive resources available. This classic example is the most famous one on the particular theme of distant passed pawn, but another, less instructive, endgame is Chigorin – Marshall, Carlsbad 1907. 35...♔f7 In the first phase both players centralise their kings. 36 ♔c1 ♔f6 37 ♔d2 ♔e5 38 ♔e3 h5 Black’s pieces are optimally placed and it’s time to advance the pawns. 39 a3 White would like to manoeuvre his knight to d3 to decentralise the opponent’s king, so 39 ♘h3! comes into mind. After the further 39...♘e6 40 ♘f2 ♘f4 Black prevents White’s manoeuvring plan with a typical device to keep the enemy knight under control. 41 ♔f3 a5 42 ♔e3 b5 43 ♔f3 ♘e6 44 ♔e3 ♘c5 45 ♔f3 White can only wait while Black improves his position move by move and this is also the best defensive strategy. 45...b4 (45...♔d4 46 ♔f4 gives White enough counterplay with his passer.) 46 ♔e3 a4 47 ♘h3! Now is the right moment to strike back. 47...♘xe4 (47...h4 48 ♘f4 ♔f6 49 ♔f3) 48 ♘f4 ♘f2 (48...♘f6 doesn’t work on account of the reply 49 ♘d3+. It’s typical that White saves himself with knight forks. Note the importance of waiting with the counterattack ♘h3 until Black has two weaknesses on h5 and b4.) 49 ♔f3 ♔f5 50 ♘xh5 ♘d3 51 ♘g3+ ♔e5 52 b3 a3 53 ♘e2 ♘c5 54 ♔e3 ♘xb3 55 ♔d3 and White’s king is close enough to draw so Black is unable to exploit his extra pawn. 39...a5 40 ♘h3 ♘c2+ 41 ♔d3 ♘e1+ 42 ♔e2 ♘g2 43 ♔f3 43 ♔d3? ♘f4+ 44 ♘xf4 ♔xf4 leads to a lost pawn ending since Black can force the white king to a square where Black will queen first with check. 45 ♔d4 h4 46 e5 ♔f5!. This is the idea. White is forced to place the king on the long light-squared diagonal and then Black gains a tempo in the pawn race after 47 ♔d5 h3 48 e6 h2 49 e7 h1♕+. 43...♘h4+ 44 ♔e3 ♘g6 45 ♘g5 ♔f6 46 ♘h7+ ♔g7 47 ♘g5 ♔f6 48 ♘h7+ ♔e7! 49 ♘g5 49 ♔d4? is a mistake as then follows 49...♘f8! and Black trades to a winning pawn ending after 50 ♘g5 ♘e6+. 49...♘e5 50 ♔d4 ♔d6 51 ♘h3 a4 52 ♘f4 h4 53 ♘h3 b6 54 ♘f4 b5 55 ♘h3 55...♘c6+?! The most precise win can be seen in the following variation: 55...♘c4! 56 ♔c3 ♔e5 57 b3 axb3 58 ♔xb3 ♔xe4 59 ♔b4 ♔d3 60 ♘g1 ♘d6! 61 ♔c5 ♘e8 62 ♔xb5 ♔e3 63 a4 ♔f2 64 ♘h3+ ♔g3 65 ♘g1 ♔g2 66 ♘e2 ♘c7+ 67 ♔c4 ♘e6 68 a5 h3 69 ♘f4+ ♘xf4 70 a6 ♘d5! 71 ♔xd5 h2 72 a7 h1♕ 73 a8♕ ♔g3+. 56 ♔e3 56 ♔c3 doesn’t save the situation after 56...♔e5 57 ♔d3 ♘a5 58 ♘g1 ♔f4 59 e5 ♘c6 60 e6 ♔f5 61 ♔e3 ♔f6! (Note that the immediate 61...♔xe6? fails to 62 ♘f3! when White makes use of the fundamental defensive resource for this type of ending to maintain material balance, i.e. knight forks!) 62 ♘f3 (62 ♔e4 ♘d8!) 62...h3 63 ♔f4 ♔xe6! 64 ♘g5+ ♔d5 65 ♘xh3 ♔c4 66 ♔e4 ♔b3 67 ♔d5 ♔xb2 68 ♔xc6 b4 and Black wins. 56...♔c5 57 ♔d3 b4! 58 axb4+ 58 ♘f4 ♘e5+ 59 ♔e2 (59 ♔c2 bxa3 60 bxa3 ♔d4) 59...b3 60 ♘e6+ ♔c4 61 ♘f4 ♔d4 62 ♘e6+ ♔xe4 63 ♘c5+ ♔f4 64 ♘xa4 h3 65 ♔f1 ♔g3 66 ♘c3 h2 67 ♘e4+ ♔f3 68 ♘f2 ♘d3 69 ♘h1 ♘xb2 and Black wins. 58...♔xb4 59 ♔c2 ♘d4+ 60 ♔b1 ♘e6 61 ♔a2 After 61 ♔c2 ♔c4 62 ♘f2 ♘g5 63 ♔d2 ♔d4 Black dominates all the white pieces. 61...♔c4 62 ♔a3 ♔d4 63 ♔xa4 ♔xe4 64 b4 ♔f3 65 b5 ♔g2 White resigned. 178 Szabó – Groszpeter Hungary 1984 White to move In this very sharp knight ending Szabó found the amazing 1 ♘d2!! ♘xd2 2 a5! and Black resigned but he should have tested White’s technique a little more since White must play precisely to win. The main variation is 2...bxa5 Or 2...♘c4 3 a6! ♘d6 4 ♔xb6 h4 5 ♔c5! h3 (or 5...♘c8 6 b6) 6 a7 ♘b7+ 7 ♔b4 h2 8 a8♕ h1♕ 9 ♕h8+ and White wins. This variation worked perfectly for White due to the fact that the black king was placed on h6. 3 b6 ♘c4 4 b7 ♘e5 179 White to move 5 ♔b8!! Such paradoxical moves are always beautiful. White obviously wants to prevent the fork on c6 by manoeuvring the king to c7 but the natural 5 ♔b6? fails to 5...♘d7+ 6 ♔c7 (or 6 ♔c6 ♘b8+ 7 ♔c7 a4 8 ♔xb8 a3) 6...♘c5! and a new fork is waiting around the corner. 5...♘c6+ One of the points of the beautiful fifth move is that 5...♘d7+ is met by 6 ♔c8! ♘b6+ 7 ♔d8. 6 ♔c7 ♘b4 7 ♔b6! Now that the knight has been lured to b4 it’s correct to play the king to b6. 7...♘d5+ 8 ♔b5 ♘c7+ 9 ♔xa5 and White wins. 180 Berger 1890 White to move 1 ♘f7! It’s amazing but 1 ♘xe8?? is only a draw since White cannot force a mate with two knights and king. One of many miracles in chess. After 1...♔h8 2 ♘g5 ♔g8 3 ♘f6+ Black plays 3...♔f8 instead of stepping into the corner with 3...♔h8?? allowing 4 ♘f7 mate. 1...♘d6 2 ♘h6+ ♔h8 3 ♘g5 …and it’s mate next move. Without Black’s knight it would have been stalemate. 181-191: Bishop Endings Bishop endings are frequently a matter of finding the right diagonals. What is characteristic though is the bishop’s inability (and ability!) to work on half the board, which is a consequence of the bishop pair being split. This inability is most clearly seen when the attacker has a bishop of opposite colour to the corner square where a rook pawn aims to be promoted against a blockading king. The balance between the bishops is most upset when bishops of opposite colour are on the board. Such endings obviously have their own rules. 181 Paulsen – Metger Nuremberg 1888 White to move In this position White must be careful that the black king doesn’t reach b7 and a8 with a theoretical draw. Correct is therefore 1 ♔d4!. In the game the natural but fatal move 1 ♔c4?? was played. A drawn position then arose after the miraculous 1...b5+! 2 axb6+ (or 2 ♔xb5 ♔b7) 2...♔b7 when White cannot win after 3 ♔b5 ♔a8!. It’s stalemate after 4 ♔c6, while 4 ♗b8 is a draw as well after 4...♔xb8 5 ♔c6 ♔a8 or 5...♔c8. Equally, 1 ♔c5?? leads to the same situation after 1...b6+!. 1...♔c6 1...b6 2 a6 prevents Black’s king from reaching b7. 2 ♗b6 ♔d7 3 ♔c5 ♔c8 4 ♗a7! White must prevent the king from reaching b8. 4...♔c7 4...b6+ is obviously met by 5 ♔xb6. 5 ♔b5 ♔d7 6 ♗b8! The key diagonal to win this ending, preventing Black from reaching the corner. 6...♔c8 7 ♗h2 ♔d7 8 ♔b6 ♔c8 9 ♗g3 White wins by zugzwang and collects the pawn next move. 182 Averbakh 1954 White to move 1 ♗e1 To be able to hold the draw the defender must prevent the pawns from reaching the fifth rank. The most effective continuation is to manoeuvre the bishop to d8 and put pressure on the weakest point in Black’s position, the g5-pawn. White can do this immediately with 1 ♗b6! followed by 2 ♗d8!. 1...f4 2 ♗f2 h4 2...♔f5 is answered by 3 ♗b6 followed by ♗d8. 3 ♔f3?? doesn’t prevent Black’s plan of placing all pawns on the fifth rank after 3...g4+ 4 ♔g2 ♔g5! 5 ♗b6 h4 6 ♗d8+ ♔h5 Black wins. 3 ♗b6! This is the last chance to manoeuvre the bishop to the important d8-h4 diagonal and avoid the winning move ...g4 in the future. Note that passive defence with 3 ♗e1? loses to 3...♔f5 (And not 3...f3+? 4 ♔g1! [4 ♔h2! works but not 4 ♔f2? h3 5 ♔g1 ♔f4 6 ♗d2+ ♔f5 7 ♗a5 ♔e4! 8 ♗c7 ♔e3 9 ♔f1 f2! 10 ♗b6+ ♔f3 11 ♗xf2 h2 and Black wins.] 4...♔f5 [4...♔f4 5 ♗d2+ ♔f5 6 ♗a5] 5 ♗a5 g4 6 ♗d8 h3 [6...g3 7 ♗xh4 ♔f4 8 ♔f1! ♔g4 9 ♗xg3] 7 ♗c7 ♔e4 8 ♔f2 with a draw.). We now have three branches: a) 4 ♗f2 ♔e4 5 ♗e1 ♔e3 6 ♗f2+ ♔e2 7 ♗c5 f3+ 8 ♔g1 h3 Black has three pawns and wins by sacrificing two of them for the bishop and then wins the pure pawn ending. b) 4 ♔h3 ♔e4 5 ♔g4 ♔e3 6 ♗a5 f3 and Black wins. c) 4 ♗a5 g4 Black wins when all pawns are placed on the fifth rank. (4...♔e4 is answered by 5 ♗d8!.) 5 ♗d8 h3+ 6 ♔h2 (6 ♔f2 g3+ 7 ♔f3 h2 8 ♔g2 f3+ 9 ♔h1 f2) 6...♔e4 7 ♗b6 ♔f3 8 ♗c7 ♔e3 9 ♗b6+ ♔e2 10 ♗c5 f3 11 ♔g1 f2+ 12 ♗xf2 h2+ 13 ♔xh2 ♔xf2 14 ♔h1 ♔g3 15 ♔g1 ♔h3 and Black wins. 3...♔f5 3...f3+ 4 ♔f2 ♔h3 5 ♔xf3 g4+ 6 ♔f2 ♔h2 7 ♗c7+ g3+ 8 ♔f3 ♔h3 9 ♗b6 g2 10 ♗g1 leads to a stalemate. 4 ♗d8! After this move there is no effective continuation since the g5pawn cannot advance any further. So White just keeps up the pressure and prevents Black’s king from improving its position. 183 Centurini 1847 White to move It’s normally easier to win with a bishop pawn or a knight pawn rather than with a central pawn. The reason for this is that the diagonals are shorter on the side of the board. According to Centurini’s rule one needs at least two available squares for the bishop to be able to hold the draw. In this example the a7-b8 diagonal is too short. 1 ♗h4 White’s main idea is to manoeuvre the bishop to b8 and force the black bishop away from the h2-b8 diagonal. 1...♔b5! Black must prevent the bishop from reaching a7. If Black plays passively with, for example, 1...♗f4 the game is lost more easily after 2 ♗f2 ♗h2 3 ♗a7 ♗g3 4 ♗b8 ♗f2 5 ♗h2 ♗a7 6 ♗g1. This position is White’s key idea, exploiting the fact that the a7-b8 diagonal is too short. Black needs an additional square to hold the draw. 2 ♗f2 ♔a6 3 ♗c5! An excellent move placing Black in zugzwang. Note that the winning try 3 ♗e3 is met by 3...♗d6 4 ♗g5 ♔b5 5 ♗d8 ♔c6 6 ♗e7 ♗h2 and now 7 ♗c5 obviously doesn’t work on account of 7...♔xc5. 3...♗e5 Of course 3...♔b5 allows 4 ♗a7. 4 ♗e7 ♔b5 5 ♗d8 ♔c6 6 ♗f6! This is the point. White gains an important tempo since Black must move the bishop while controlling the promotion square. Black doesn’t have the extra move he needs to manoeuvre the king to a6 and stop White’s ♗a7. 6...♗f4 7 ♗d4 ♗g3 8 ♗a7 ♗f4 9 ♗b8 ♗e3 10 ♗g3 ♗a7 11 ♗f2 White’s subtle manoeuvring is at last rewarded by reaching this winning position. 184 Averbakh 1977 White to move There are four cases which have been analysed in depth by Centurini and Averbakh. The basic winning position, which is the First Case, is a situation where the defender cannot control the squares in front of the pawn. Positions where the defender stands close to the pawn but not in front of it makes it easier for the attacker to exchange bishops. In this position White wins by exchanging bishops either on c6 or d7. White wins after 1 ♗f3! 1 ♗g4 and 2 ♗d7 takes a longer time since the bishop gets in the way of its own pawn. 1...♗a4 2 ♗c6 ♗xc6 3 ♔xc6 ♔e8 4 ♔c7 and the pawn promotes. 185 Kurajica – Markland Hastings 1971/72 Black to move The main question to answer when facing this type of ending is whether the defending king should manoeuvre behind the pawn or in front of it. The first method is called the Second Case and the second method the Fourth Case. The Third Case by the way means that the defending king is behind the pawn but one of the diagonals is less than four squares long. In the game 1...♔d7?! was played. However, the safest defence is to go behind the pawn rather than to place the king in front of it, unless it can reach the f7-square by force. After 1...♔d5! 2 ♔g6 ♔e4 3 ♗c7 ♔f3 4 ♗d8 ♔g4 Black holds the draw by achieving rear opposition. This is a very important idea to remember and we’ll elaborate more on it in the next position. 2 ♔g6 ♔e8! If Black cannot reach rear opposition it’s important to reach front opposition. 3 ♗f4 3 ♗d6 prevents the black king from reaching f8 and g8 but if Black keeps control of the long dark-squared diagonal with 3...♗d4 the draw is secured nevertheless. 3...♗d4? 3...♔f8! 4 ♗h6+ ♔g8 is a draw since White cannot play the bishop to g7. If White places the bishop on f6 it’s still a draw as the following variation shows: 5 ♗g5 ♗d4 6 ♗f6 ♗f2 7 ♗e5 ♗h4 8 ♗f4 ♗d8 9 ♗g5 ♗xg5! 10 ♔xg5 ♔g7. The conclusion is that front opposition works if the enemy pawn hasn’t reached the sixth rank. 4 ♗h6! It’s not possible to prevent ♗g7 next move so Black resigned. 186 Capablanca – Janowski New York 1916 Black to move Here is the most famous position with this particular distribution of material. What is incredible is that the same position arose in another high level game between Taimanov and Fischer at Buenos Aires in 1960. The only slight difference was that Black’s king was placed on f4 with White to move. Capablanca played 83 ♔d5 and Janowski resigned in a theoretically drawn position. According to Averbakh’s analysis he should have continued 83...♔f4! 84 ♗d4 84 ♗e5+ makes it easier for Black to go behind the pawn by 84...♔e3 85 b5 ♔d3 86 ♔c6 ♔c4. 84...♔f3! Black must go behind the pawn. 85 b5 After 85 ♗c5 ♔e2 86 ♔c6 ♔d3 87 ♔d7 ♗g5 88 b5 ♔c4 it’s also a draw. 85...♔e2 86 ♔c6 ♔d3 87 ♗b6 ♗g5 88 ♔b7 88 ♗c7 ♗e3 89 ♗d6 ♔c4 makes no difference. 88...♔c4 89 ♔a6 ♔b3! 90 ♗f2 ♗d8 91 ♗e1 ♔a4!. The manoeuvre ♔c4-b3-a4 is very important to remember. It is the key to holding the draw when defending with the help of rear opposition. Taimanov played 82 b5 and Fischer drew after 82...♔e4 83 ♗d4 ♗c7 84 ♔c5 ♔d3! 85 ♔c6 ♔c4! (rear opposition) 86 ♗b6 ♗g3 87 ♗a7 ♗c7!. This was the only game Taimanov managed to draw against one of the greatest players in history and apparently Fischer knew his predecessors very well indeed. In other words, this important position was already stored in his head. 187 Illescas Cordoba – Tukmakov 1993 White to move Four pawns against three on the same wing is normally an easy draw. But if the defender has a weakened pawn structure there are some practical winning chances. In this endgame one grandmaster managed to fool another. 39 ♔g2 ♔f7 40 ♔f3 g6 Black follows the correct strategy which is to place the pawns on the opposite colour of its own (and the opponent’s!) bishop. 41 ♔e4 ♗d6 42 ♗d4 ♗c7 43 h3 White’s pieces are placed on the best squares so it’s time to advance the pawns. 43...♗d6 44 g4 hxg4 45 hxg4 ♗c7 46 ♗e5 ♗b6 47 f3 White places the pawns on the opposite colour of his bishop. 47...♗c5 48 ♔d3 White would like to manoeuvre the king to d7. 48...♗b6 49 ♔c4 ♗e3 50 ♗d4 ♗f4 51 ♔b5 ♗g3 52 ♔c6 ♗f4 53 ♗c5 ♗g3 54 ♗d6 ♗f2 55 ♗e5 ♗e3 56 ♔d7 56...♗d2? The Hungarian GM Pintér in 1000 Minor Piece Endings recommends 56...g5 57 ♔c6 ♔e7 58 ♔b5 ♔f7 59 ♔c4 ♔g6 60 ♔d3 ♗c5 61 e3 ♗b6 62 f4 gxf4 63 exf4 ♗c5 64 ♔e4 ♗e7 with a draw. 57 e4? 57 f4 should have been played immediately since 57...g5 is met by 58 f5 giving White two passed pawns. 57...♗e3? 57...g5 was necessary in order to stop f3-f4. 58 ♗d6? Again 58 f4 should have been played. 58...♗d2? The last opportunity to execute 58...g5! is gone. 59 f4! It’s interesting that none of the players understood the importance of this pawn push but of course time pressure might have been involved. 59...♗c3 59...♗e3 would lose after the spectacular 60 f5! gxf5 61 gxf5 exf5 62 e5 ♗g5 63 e6+ ♔g8 64 ♗f4 ♗f6 65 ♗c7 f4 66 ♗d8!. The slower 66 ♗xf4 also won. It’s useful to remember this pawn breakthrough which commenced with 60 f5!. An extra pawn can always be exchanged for a positional advantage. Quantity is exchanged for quality. In this case the passed e-pawn was more dangerous than Black’s passed f-pawn. 60 e5! Black resigned since it’s not possible to prevent the decisive f4-f5 after which the e-pawn decides. An instructive example which shows that both players had problems realising not only the value but also the right moment when to place a pawn on the same colour as the bishop. Black had several chances to play ...g5 until White played his pawn to f4. 188 Wojtkiewicz – Khalifman Rakvere 1993 White to move Capablanca formulated the famous and important rule “When the opponent has a bishop, keep your pawns on squares on the same colour as this bishop. But if you have the bishop, then irrespective of whether or not the opponent has a bishop, keep your pawns on squares of the opposite colour to that of your bishop.” However, there are exceptions to Capablanca’s rule. It’s correct to place the pawns on the same colour as the opponent’s bishop if it’s not possible to attack them. In this position it’s important to fix Black’s queenside, thereby making it a future target for the white pieces, so Wojtkiewicz’s move 30 a4! makes a lot of sense. Note that the routine move 30 ♔d4? is met by an activation of the doubled pawns with 30...b5! followed by ...b6. 30...g5 31 ♔d4 ♗f7 32 ♗f3 White prevents Black from playing ...h5 since it’s in Black’s interest to exchange pawns. 32...♗e6 33 f5! By placing pawns on white squares the black bishop becomes less and less effective. 33...♗f7 34 b4 ♗e8 35 b5 White has five pawns on white squares and only one of them can be moved without being lost next move. It’s rare that bishop endings are treated in this manner but in the present position this is the correct procedure. 35...♗f7 36 ♗d1 ♗g8 37 ♗b3 ♗f7 38 e4 After this decisive pawn move, exploiting the pin on the classical diagonal, all white pawns are placed on white squares! Capablanca’s general rule is thus completely irrelevant here. 38...♗g8 39 ♗a2 ♗f7 40 ♗xd5 The b7-pawn is a target so Black has to exchange bishops. 40...♗xd5 41 exd5 A very interesting pawn ending has arisen. 41...♔e7 42 ♔c3! White plans ♔b4 followed by a5. 42...♔d6 43 ♔c4 43...♔e5 Black has not managed to stop the decisive ♔b4 as the following forced variations show: 43...♔d7 44 ♔b4 ♔d6 45 a5 bxa5+ (45...♔xd5 46 a6 bxa6 47 bxa6 ♔c6 48 ♔a4!! is a pretty variation.) 46 ♔xa5 ♔xd5 47 ♔b6 ♔c4 48 ♔xb7 ♔xb5 49 ♔c7 ♔c5 50 ♔d7 ♔d5 51 ♔e7 ♔e5 52 ♔f7 and White wins the f6-pawn. 44 a5! This tactical breakthrough secures the c5-square for the king so it can escort the d-pawn to d6. 44...bxa5 45 ♔c5 a4 46 d6 b6+ 47 ♔c6 a3 48 d7 a2 49 d8♕ a1♕ A third endgame has arisen! The queen ending is easily winning for White due to his advanced position. 50 ♕d6+ The immediate capture 50 ♕xb6 was also good. 50...♔e4 51 ♔xb6 ♔f3 52 ♔b7 ♔g2 53 ♕d3 White plays in a simple and pragmatic manner. A quicker win was 53 b6 ♔xh3 54 ♔c7. 53...♕c1 54 b6 ♕c5 55 ♕b3 ♔h2 56 ♕f3 ♕d4 57 ♕c6! ♔xh3 58 ♔c8 ♕b4 59 b7 ♕f8+ 60 ♔d7 ♔xg4 61 ♕c8 Black resigned. 189 Averbakh 1954 White to move If in the previous example the attacker exploited his positional advantage by placing his pawns on the same colour as the bishop and going against the famous rule by Capablanca, then in contrast this study by Averbakh follows the rule by Capablanca that states: “If you have a bishop, then, irrespective or whether or not the opponent has a bishop, keep your pawns on squares of the opposite colour to that of your bishop.” I have been familiar with this study ever since I was a teenager and the reason was that my friend Peter Fransson showed it to me several times and was fascinated by it. In fact he was so fascinated that he specialised in this type of bishop ending and produced instructive material on it. Averbakh has written that virtually all bishop endings depend on the availability of invasion squares. In this position Black’s king guards against an invasion by the white king, so the question is how can White cede the move to Black. Obviously the f3-square corresponds to the f7-square so the trick is to lure Black, who has less space on the diagonals, to lose a move. Averbakh’s solution runs as follows: 1 ♗e2 ♗g6! The most stubborn defence. If 1...♗e8 2 ♗d3 ♗g6 (2...♗d7 3 ♗c2 ♗e6 4 ♗d1 ♗f7 5 ♗f3) 3 ♗c2 ♗h7 4 b3! ♗g8 5 ♗d1 ♗f7 6 ♗f3. 2 ♗d3 ♗h7 3 ♗c2 Dvoretsky suggests the simpler 3 ♗f1! ♗g6 (3...♗g8 4 ♗e2 ♗f7 5 ♗f3) 4 ♗g2 ♗f7 5 ♗f3. 3...♗g6 4 ♗b1!! Averbakh explains that Black has only two squares available on the b1-h7 diagonal, whereas White has three. This is the reason Black eventually loses the battle. Each square for White corresponds with a square for Black. 4...♗h7 5 ♗d3! The h7-square corresponds with the d3-square. 5...♗g6 6 ♗c2! ♗h7 7 ♗b3!! ♗g8 8 ♗d1 ♗f7 9 ♗f3 White has eventually exploited his advantage with the help of zugzwang and Black is completely lost. 190 Vaganian – Karpov Leningrad 1969 White to move One of the leading principles to play for a win in endings with bishops of opposite colour is to create two passed pawns. In this position the main question to answer is whether this is possible. Karpov, who is famous for his skill in handling bishops of opposite colour positions, played 48...e4! The e-pawn in itself is not worth anything. More important is the space it vacates for the king. 49 fxe4 49 g4 hxg4 (49...h4 was playable too.) 50 hxg4 e3 and Black has two passers to secure the win, as can be seen after 51 ♔d3 ♔e5 52 ♗b3 ♔d6 53 ♗a2 ♔c5 54 ♗g8 ♔b4 55 ♗a2 ♗e5 56 ♗f7 e2 57 ♔xe2 ♔c3 and Black breaks through. 49...g4! Fixing the weakness on g3. 50 hxg4 hxg4 51 ♗b3 ♗e5! Forcing White to defend the pawn with his king. 52 ♔f2 ♔e7 53 ♔e3 Otherwise Black eventually wins the bishop. 53...♗xg3 54 ♔d4 ♗e5+ White resigned. 191 Engqvist – Agrest Osterskar 1994 White to move The trick to win this position is to manoeuvre the bishop to the right diagonal. 66 ♗e1! 66 e6+?? ♗xe6 57 fxe6+ ♔xe6 is obviously a draw since it’s impossible to stop the king from entering the a8-corner which is a well-known theoretical draw. 66...♗d7 67 ♗g3! ♗c8 68 e6+ ...and Black resigned due to 68...♗xe6 69 fxe6+ ♔xe6 70 ♔e4 ♔d7 71 ♔d5 ♔c8 72 ♔c6 and Black’s king cannot reach the corner. 192-205: Knight and Bishop Endings The short-range knight versus the long-range bishop introduces among other things the theme of domination, where the bishop sometimes dominates the knight or the other way around. However, what is striking is that in this particular ending the pawn position is almost decisive in determining which of the two pieces is the strongest. Apart from these features we’ll also look at different methods how to mate with knight and bishop versus a lone king. The section ends with some positions where the bishop pair show their endgame talents against knight(s). 192 Fischer – Taimanov Candidates match, Vancouver 1970 Black to move Taimanov played the dire 81...♔e4?? placing the king on a white square instead of: a) 81...♔d4 82 ♗e2 ♘d7+ 83 ♔g7 ♘c5 84 h4 ♘e4 85 h5 ♘g3 86 h6 ♘f5+ or b) 81...♔d6 82 ♗e2 ♘d7+ 83 ♔g7 ♘b6 84 h4 ♘d5 85 h5 ♔e7 86 h6 ♘f6 87 ♗f3 ♘e8+ with an easy draw. An amusing episode is that Botvinnik heard about the position on the phone in Moscow and instantly, without a chess set, realised that Taimanov had a more spectacular draw with c) 81...♘d3! 82 h4 ♘f4 83 ♔f5 ♔d6! 84 ♔xf4 ♔e7 taking advantage of the fact that the white-squared bishop cannot win with the wrong rook’s pawn, since the king reaches the dark-squared corner. Now Fischer was able to exploit Black’s unfortunate king’s position ...by 82 ♗c8! Taimanov must have forgotten about this long bishop move. 82...♔f4 After 82...♘f3 83 ♗b7+ the check on the long white diagonal decides. 83 h4 ♘f3 84 h5 ♘g5 85 ♗f5 ♘f3 86 h6 ♘g5 87 ♔g6 Black is in zugzwang and must move his knight. 87...♘f3 88 h7 ♘e5+ 89 ♔f6 Black resigned. 193 Stein – Dorfmann USSR 1973 White to move This position is a draw since after 1 ♗f6 ♘d3 2 ♗a1! ♘b2 3 ♔e1 ♔b1 4 ♔d2 ♔xa1 4...♘c4+ 5 ♔d1 ♔xa1 makes no difference after 6 ♔c2!. 194 White to move The knight is a peculiar piece since it is unable to gain a tempo so it’s impossible for White to force Black to move. To understand this position it’s simplest to think that the knight controls different coloured squares on each move and that’s the reason this position is a draw. When the knight and the defender’s king are placed on the same colour it’s a draw if the attacker is to move. Here White can exploit the fact that the knight can never gain a tempo by placing the king on the same colour as the knight. 5 ♔c1! 5 ♔c2?? ♘d3!. 5…♘c4 6 ♔c2 and the players agreed to a draw since it’s impossible to achieve a zugzwang. An extraordinary position and one of many miracles in chess. 195 Fransson – Engqvist Lulea 1980 Black to move At the time Bjorn Borg played the Wimbledon final against John McEnroe I reached this interesting ending in the Swedish Junior Championship. 77...♗c3 All reasonable moves along the long diagonal except one lead to a draw. 77...♗e5?? loses after 78 ♘d8+ ♔g7. Necessary since the decisive f5-f6+ must be prevented, but it doesn’t help after 79 h8♕+! ♔xh8 80 ♘f7+. The trick to draw this ending is to avoid forks and place the bishop far away from the knight. 78 ♔h6 ♗h8 79 ♔g5 ♗c3 80 ♔g4 ♗b2 81 ♔f3 ♗c3 82 ♔e4 ♗f6?? The decisive mistake giving White the opportunity of a fork. 82...♗b2 or 82...♗a1 would leave a positional draw. 83 ♘f4! ♔g7 The only way to prevent ♘g6... ...but not the fork 84 ♘h5+ after which I resigned. I still remember that right after the game IM Konstanty Kaiszauri told me I had missed a theoretical draw. This episode provides some proof that this position was one of the important positions in Kaiszauri’s head, since he knew the correct evaluation of it way ahead the Tablebase we have available online today. According to Encyclopaedia of Chess Endings – Minor Piece Endings I’m not the only victim in this ending. Mikhalchisin lost to Anikaev in the USSR 1981 and Helgi Olafsson lost to Ivanchuk at Reykjavik 1990. Such defeats in theoretically drawn endings just go to show that this ending is tricky in practice. However, the first thing you have to know is that the position is a draw and I didn’t in 1980. Note that this ending is a draw even if the pawns are placed further back. However, this is a rare ending, a once in a lifetime event, and it really impressed me that Kaiszauri had this knowledge, unlike me. However, I now understood how important it was to acquire this knowledge since I had just lost a drawn endgame. 196 Flohr – Capablanca Moscow 1935 Black to move In this endgame White has a slight advantage but Black should be able to hold if he places his pawns on b6 and a5 according to Capablanca’s rule. 23...♔e7 24 ♔d2 ♔d6 25 ♔c3 b6! 26 f4 ♗d7 27 ♘f3 White’s king belongs on d4 and so the knight manoeuvres to a square from where it can exert pressure on the isolated pawn. 27...f6! 28 ♔d4 a5! Capablanca has followed his own advice and placed pawns on the opposite colour of his bishop. 29 ♘d2 ♗c8 30 ♘b1 ♗e6 31 ♘c3 ♔c6 32 a3 h6 33 g3 33...h5! Capablanca now breaks his own rule – because he does not want to allow his opponent a free hand on the kingside. If Black remains completely passive White can considerably improve his position by playing his knight to h4 followed by f4-f5 and g3-g4 and then manoeuvre the knight to e6 via g2 and f4 and win when the king is placed on c6. White’s h-pawn can help to set up such a zugzwang. Of course Black’s bishop has options of disturbing the manoeuvring ability of the knight while his h-pawn can help to prevent future expansion with g3-g4. It’s not at all easy for White to get at the pawn on h5 and so it should be okay to place it there. Another continuation was to follow Capablanca’s advice and let the h-pawn stay on h6 and instead play 33...♔d6!? and keep active with his bishop after 34 ♘d1 (Continuations like 34 f5 ♗xf5 35 ♘xd5 b5 or 34 g4 ♗xg4 35 ♘xd5 b5 are only in Black’s interest as he is then rid of the isolani.) by 34...♗g4 35 ♘f2 ♗e2! with excellent drawing chances. If White places his kingside pawns on white squares they will be targets for the bishop to attack from the rear. So there are actually two defensive strategies to adopt in this type of ending. It’s good to know that there is room for several strategies, since it means the margins for holding the draw are pretty good. If the diagram position is the worst you can get when playing with an isolani, it means it’s still acceptable. 34 b4 axb4 35 axb4 ♔d6 36 b5! Fixing b6 so from now on Black must be more careful when calculating concrete variations. 36...g6 A careless move here, for example, would be 36...♗f7? as can be seen after 37 f5 ♗g8 38 ♘e2 ♗f7 39 ♘f4 ♗e8 40 ♘xd5 ♗xb5 41 ♘xb6. 37 ♘a4 ♔c7 38 ♘c3 ♔d6 39 f5! The only way to play for a win. 39...gxf5 Black is forced to disrupt his pawn formation since 39...♗xf5 fails to 40 ♘xd5 ♗d3 41 ♘xf6 ♗xb5 42 ♘d5 ♔c6 43 ♘e7+ and the g6-pawn is lost. 40 ♘e2 ♗d7 Averbakh recommends 40...♗g8! 41 ♘f4 ♗f7 42 h3 ♗e8 43 ♘xd5 ♗xb5 44 ♘xb6 ♗c6 but Capablanca’s choice is good enough. 41 ♘f4 ♗e8 41...♗xb5? 42 ♘xh5 is only risky for Black. 42 ♘xd5 ♗xb5 43 ♘xb6 ♗c6 44 ♘c4+ ♔e6 45 ♘b2 ♗b5 46 ♘d1 ♗e2 47 ♘f2 ♗f1 There is no reason to allow ♘h3-f4 even though it isn’t the end of the world. 48 ♘d3 ♗xd3 49 ♔xd3 197 Black to move A very interesting pawn endgame has arisen. 49...♔e5! Another draw was possible after 49...♔f7 but naturally the centralising move is more appetizing because it limits White’s choices. 49...♔d5? loses after 50 ♔d2!! as the following subvariations show: a) 50...♔e4 51 ♔e2 ♔d5 (or 51...h4 52 gxh4 f4 53 h5 ♔f5 54 exf4) 52 ♔f3 ♔e5 53 h3 ♔d5 54 ♔f4 ♔e6 55 h4. b) 50...h4 51 gxh4 f4 52 exf4 ♔e4 53 h5 ♔f5 54 ♔e3. c) 50...♔e5 51 ♔e1!! ♔d5 (51...♔e4 52 ♔e2 wins.) 52 ♔f2! ♔e4 53 ♔e2. These variations turn once again to the subject of corresponding squares which we have touched on before. 50 ♔e2 50 ♔d2 is met by the strong 50...h4! which is the only move to draw. (50...♔e4? loses after 51 ♔e2.) 51 gxh4 f4! 52 h5 fxe3+ 53 ♔xe3 ♔e6 with a draw. 50...♔e4 The only move. 51 h3 51 ♔f2 h4! draws as we saw in the previous variation. 51...♔d5 According to many sources this is the only way to draw since 51...♔e5? 52 ♔f3 ♔d5 53 ♔f4 ♔e6 54 h4 as well as 51...h4? 52 gxh4 f4 53 h5 lead to a win. However, the super move 51...f4!! holds the draw too! After 52 gxf4 (52 exf4 h4 53 ♔f2 f5! 54 ♔g2 ♔e3 55 ♔h2 ♔f3 and White is forced to capture the h-pawn with an easy draw for Black.) 52...h4 53 ♔f2 ♔d3 54 ♔f3 f5 55 ♔f2 ♔d2 and it’s a positional draw. 52 ♔f3 ♔e5 After Capablanca made the only move (You can check why it’s the only move!) the great endgame connoisseurs agreed to a draw. 198 Kudrin – Gulko USA Championship 1988 White to move In his excellent book Winning Pawn Structures – The Isolated Queen’s Pawn and its Derivatives GM Alexander Baburin thinks that, compared with the previous example, White’s pawn formation is somewhat better. One reason is that it’s more difficult for Black to play ...a7-a5 because of White’s possibility of responding b2-b4. If White captures on b4 White recaptures with the c-pawn, thereby obtaining a pawn majority with chances of later establishing an outside passed pawn. 31 a3 A good rule to follow is to place all pieces on the opposite colour of the enemy bishop. 31...♗e6 32 ♔e3 ♔d6 33 ♔d4 h6 34 ♘f4 ♗f7 35 ♘d3 Baburin thinks that one possible plan for White is to advance the f-pawn to f5 followed by ♘f4 with a grip on the kingside. This explains Black’s reaction in the game. However, another idea might be the long-term plan b4-b5 followed by the knight manoeuvre ♘b4-c6xa7. 35...g5 This is Black’s best set-up on the kingside. 36 g3 ♗h5 37 ♔e3 A more direct approach was Baburin’s recommendation 37 f4 gxh4 38 gxh4 followed by a knight manoeuvre to e3 where it can head for one of the three squares d5, f5 or g4. Black has three weaknesses in the centre and on the kingside, while a7 might be a fourth weakness. 37...♗g6 38 ♘f2 ♗f5? 38...♗h5, targeting the f3-pawn, would have made White’s task more difficult. In such a case White could have changed the plan to the one recommended by Baburin on move 37. 39 ♔d4 ♗e6 40 ♘d1 ♗d7 41 ♘e3 White plays like Flohr did in the previous endgame by placing the king on d4 and the knight in a position where it can attack the isolani. This is the correct deployment of the pieces before deciding where to place the pawns. 41...♗e6 42 b3! White has all the time in the world to play as he pleases and in an Ulf Andersson manner. Such a nice quiet move is representive of no-hurry chess which is of great significance, technically as well as psychologically, in advantageous endings like this one. 42...♔c6 43 b4! ♔d6 44 b5! If White should be able to win this position it’s not enough to do so with the weakness on d5. White needs to create another weakness and this has now been accomplished. White has exploited the fact that Black didn’t move his a-pawn and it’s now fixed on a7. There it’s a potential target which can be attacked by the white knight on the outpost c6. 44...h5 45 f4 gxf4 46 gxf4 ♗g8 47 ♘f5+ ♔d7 48 ♘g3 ♗f7 49 f5! Black’s position has collapsed and the reason was the two strong pawns on the fifth rank! 49...♔e7 Or 49...♔d6 50 ♘e2 ♗e8 51 ♘f4 ♗xb5 52 ♘xh5 and White wins. 50 ♘e2 ♗e8 51 ♘f4 ♗d7 52 ♘xh5 ♗xf5 53 ♔xd5 Good was also 53 ♘f4. 53...♗d7 54 c4 ♗e6+ 55 ♔d4 ♗f5 56 ♘f4 ♔d6 57 h5 ♗b1 58 h6 ♔e7 59 c5 bxc5+ 60 ♔xc5 ♗c2 61 a4 ♔f7 62 a5 ♗e4 63 ♔d6 ♔g8 64 ♘d5 ♗d3 65 b6 axb6 66 axb6 ♗a6 67 ♔c7 Black resigned. It’s very interesting and kind of magical that there is such a big difference having a pawn on e3 and c3. This knowledge can of course be exploited by striving for this pawn structure and starting to exchange the right pieces. The knowledge of important positions like this one automatically gives sustenance for winning formulae. 199 Botvinnik – Tal World Championship match, Moscow 1961 White to move In this situation the knight is much stronger than the bishop due to Black’s pawn weaknesses which will be attacked by the knight. 31 ♘d4 White attacks the c6-pawn by threatening e2-e4. 31...c5 32 bxc5 bxc5 33 ♘b5 Now the a7-pawn is under attack. 33...a6 33...a5 is answered by 34 ♘c3 ♗c4 (Or 34...♗c6 35 ♔d2 according to Botvinnik.) 35 ♘e4 or 35 ♔e1!? with the idea of ♘a4. 34 ♘c7 ♗c4 35 ♘e8 And now the f6-pawn is attacked. Note that so far the knight has attacked pawns on dark-coloured squares but after Black’s next move the attack will shift to light ones. This is the main reason the knight is actually better than the bishop when hunting pawns, since the bishop is limited to one specific colour. 35...f5 36 h4! White places the pawn on the opposite colour of the opponent’s bishop. There is no reason to allow 36 ♘f6+ ♔g7 37 ♘d7 ♗f1. 36...♔f8 37 ♘d6 37 ♘f6? ♔e7 38 ♘xh7?? f6 is obviously the wrong pawn to grab! 37...♗f1 38 g3 Note that almost all the white pawns are placed on dark squares to escape the attention of the bishop. 38...♔e7 39 ♘xf5+ ♔e6 40 e4 ♔e5 41 ♔d2 Black resigned. 200 Stoltz – Kashdan Olympiad, The Hague 1928 Black to move This position is a classical example of how to play in an open position with a bishop versus a knight. The first question to ask is whether the knight dominates the bishop or is it the bishop that dominates the knight? Other important issues in this position concern who has the most active king and the better pawn structure. When the owner of the bishop has the more active king it can prove decisive when there is a distribution of pawns on both sides of the board. However, this particular position should end in a draw with correct play. First the players activate their kings. 28...♔f8 29 ♔f1 ♔e7 30 ♔e2 ♔d6 31 ♔d3 ♔d5 The kings are now placed on their most active squares in the middle of the board and Black’s is obviously the more active of the two. 32 h4 ♗c8 This move constitutes one of the main advantages with a bishop versus knight in an open position. Black plans ...♗a6, thereby forcing the king to make a decision. 33 ♘f3? More accurate was 33 f3. Averbakh gives 33...♗a6+ 34 ♔e3 ♔c5 35 ♘c2 ♗f1 36 g3 ♗a6 37 ♘d4 ♗b7 38 ♔d3 ♔b4 39 ♔c2 ♗d5 40 ♔b2 g6 41 ♔c2 a6 42 ♔b2 and it’s not clear how Black should proceed. 33...♗a6+ 34 ♔c3 A variation like 34 ♔e3 ♔c5 35 ♘g5? ♔b4 36 ♘xf7 ♔xb3 should be avoided since it’s clearly in Black’s favour due to the passed pawn supported by the king and bishop. The bishop, which is a long-range piece, has the advantage that it can support its own pawn while halting the advance of the opponent’s. This is something the short-range knight cannot do. 34...h6 35 ♘d4 g6 36 ♘c2? 36 f3 was more useful – to control the important e4-square. 36...♔e4! The most testing move. 37 ♘e3 f5 37...h5!, planning ...f6 and ...g5, was more dangerous for White but the move played is good enough. 38 ♔d2 f4 39 ♘g4 39 ♘c2 loses to 39...♗f1 40 ♘e1 ♔f5 41 f3 g5 42 hxg5 ♔xg5 and Black runs to g3, picking up the g2-pawn. 39 f3+ ♔d4 40 ♘c2+ ♔e5 41 ♘b4 ♗b5 is also not tenable in the long run for White. 39...h5 40 ♘f6+ ♔f5 41 ♘d7 ♗c8 42 ♘f8 42 ♘c5 is answered by 42...♔g4!. 42...g5 43 g3 43 hxg5 ♔xg5 and the knight will be picked up in three moves. 43...gxh4 44 gxh4 ♔g4 45 ♘g6 ♗f5 46 ♘e7 ♗e6 47 b4 ♔xh4 48 ♔d3 ♔g4 49 ♔e4 h4 50 ♘c6 ♗f5+ 51 ♔d5 f3! An important move liberating the f4-square for the king. If 51...h3 then of course 52 ♘e5+. 52 b5 h3 53 ♘xa7 h2 54 b6 h1♕ 55 ♘c6 ♕b1 56 ♔c5 ♗e4 White resigned. 201 Capablanca A Primer of Chess 1935 White to move I’ve heard stories about 2200-players who failed to win this ending. It will probably happen once or twice in your life that you have to play this ending as an attacker or as a defender. However, the most important thing is not the ending in itself. It’s just the means to learn the method how to coordinate king, knight and bishop and develop a deeper feeling for the pieces’ abilities. I used to teach children this endgame at a very early stage and they should know how to do it before they reach an Elo rating of 1500 to get a good feeling for how knight and bishop can harmonise together like a bishop pair. There are three different methods to checkmate the lone king and we’ll look at each of them in turn. All three methods must have the objective of systematically driving the opponent’s king into the corner of the board that is of the same colour as the bishop. Capablanca’s method was the first I learned as a junior and it involves bringing up all the pieces to drive the king down the board. This is how to do it: 1 ♔e2 ♔d7 2 ♔d3 ♔e6 3 ♗f4 ♔d5 4 ♘e2 ♔c5 5 ♘c3 The first step to driving the king away from the little centre (e4d4-d5-e5) is accomplished. 5...♔b4 5...♔c6 6 ♔c4 would have accomplished the second step which is to drive the king away from the large centre f3-c3-c6-f6. 6 ♔d4 The third step is to drive the king to the last rank. 6...♔a5 The defender should head for one of the light-squared corners. 6...♔b3?! leads to a much quicker mate. 7 ♔c5 ♔a6 8 ♔c6 ♔a7 9 ♘d5 ♔a8 10 ♘b6+ ♔a7 11 ♗c7 ♔a6 12 ♗b8 This is the main set-up to learn by heart because it will be repeated according to a pattern involving a V-manoeuvre with the knight as well as with the bishop, while the king stands in opposition. 202 Black to move 12...♔a5 13 ♘d5 ♔a4 Black conducts an active defence, planning to run to the other white square on h1. Passive defence after 13...♔a6 shows very clearly the V-manoeuvre with knight and bishop. 14 ♘b4+ ends the manoeuvre from b6 via d5 to b4 which has the shape of a V. After 14...♔a5 the white king follows the king by frontal opposition: 15 ♔c5 ♔a4 16 ♔c4 ♔a5 and now it’s time to execute the Vmanoeuvre with the bishop: 17 ♗c7+ ♔a4 18 ♗b6. Compare this position with the diagram and notice that all pieces are placed according to the same pattern. Everything will be repeated one more time: 18...♔a3 19 ♘d3 ♔a4 20 ♘b2+ ♔a3 21 ♔c3 ♔a2 22 ♔c2 ♔a3 23 ♗c5+ ♔a2 24 ♗b4 ♔a1. Now it’s mate in three moves after 25 ♘d3 ♔a2 26 ♘c1+ ♔a1 27 ♗c3 mate. 14 ♔c5 14 ♗a7 ♔b3 (14...♔a5 15 ♗b6+ ♔a4 16 ♔c5 ♔b3 17 ♗a5) 15 ♔b5 ♔c2 16 ♘f4 ♔d2 17 ♗f2 is how to drive the king to the corner according to Chéron’s method which we will discuss in the next position in more detail. 14...♔b3 15 ♘b4! This knight move must be learned by heart. It’s important to control two white squares at the same time, in this case c2 and d3. 14...♔c3 16 ♗f4! White first controls the d2-square. 16...♔b3 17 ♗e5! This is stronger than 17 ♗d2 because it controls both c3 and b2. 17...♔a4 18 ♔c4 We are back to the passive defence and the V-manoeuvres. 18...♔a5 19 ♗c7+ ♔a4 20 ♗b6 ♔a3 21 ♘d3 ♔a4 22 ♘b2+ ♔a3 23 ♔c3 ♔a2 24 ♔c2 ♔a3 25 ♗c5+ ♔a2 26 ♗b4 ♔a1 27 ♘d3 ♔a2 28 ♘c1+ ♔a1 29 ♗c3 mate. The above mentioned method inspired by Capablanca’s A Primer of Chess is my preferred procedure when teaching this ending, because it’s the easiest to learn by heart. There are two other methods which are good to know about since it will increase your ability to play with knight and bishop in all kinds of situations on the board. If the first method could be baptised Capablanca’s method, meaning that the activation of the king is the first thing to do, then the second method could be called Euwe’s or Karpov’s method. In his endgame book Euwe presents a method where the knight is placed on one of the central squares and then defended by a bishop from a distance, for example a fianchettoed bishop on g2 defends a knight on d5. The advantage of this set-up is that wherever the king will try to escape, the knight will be in the right position. The reason Karpov is named in this method is that, contrary to Capablanca, he realised that in endgames piece centralisation is often more important than king centralisation. Of course in this ending it’s just a matter of taste. Now, when you are aware of the second method where centralisation of the pieces goes before the centralisation of the king, we go to what is probably the most difficult method which effectively features the Pythagorus theorem of the hypotenuse. 203 Chéron Lehr-und Handbuch der Endspiele Band 2 1964 White to move If we think of the b1-h7 diagonal as the hypotenuse of a triangle, White’s mission is to make the triangle smaller and smaller by changing diagonals from b1-h7 to d1-h5 and eventually from d1-h5 to f1-h3 when mate comes quickly. Chéron presents the following variation: 1 ♗c2 First White controls the b1-h7 diagonal. 1...♔e3 2 ♔c1!! ♔e2 3 ♗g6 ♔e3 4 ♔d1 The point of 2 ♔c1!!. The king is not needed in the centre since the minor pieces prevent the king from going there anyway. 4...♔f2 4...♔f3 5 ♔d2 ♔f2 shows the moment when to change diagonal with 6 ♗h5. 5 ♔d2 ♔f3 6 ♔d3 ♔g4 7 ♔e3 ♔h4 8 ♔f4 ♔h3 9 ♗h5 The bishop switches to the shorter d1-h5 diagonal and Black’s king is imprisoned in an even smaller triangle. Of course 9 ♗e4 mates two moves quicker but the aim here is to present Chéron’s hypotenuse method, rather than the quickest computer mate. Humans need human methods. 9...♔g2 10 ♔e3 ♔g3 11 ♗e2 Here the bishop controls the f1-square, preparing for a later ♗f1 controlling the f1-h3-diagonal. 11...♔h4 12 ♔f4 ♔h3 13 ♘g5+ ♔g2 14 ♘e4 ♔h3 15 ♔g5 ♔g2 16 ♔g4 ♔h2 White mates in six moves: 17 ♗f1! ♔g1 18 ♗h3 ♔h2 19 ♘c3 ♔g1 20 ♔g3 ♔h1 21 ♗g2+ ♔g1 22 ♘e2 mate. In this variation the king was mated on a black square next to the corner. However, if instead 19...♔h1 the king would have been mated in three moves by the bishop in the corner square itself. It’s definitely worthwhile to learn the hypotenuse method but it demands precise play with the king. My recommendation is that you learn all three methods and then choose the most appropriate one depending on the position. 204 Botvinnik – Tal World Championship match, Moscow 1961 Black to move According to Mega Database 2018 this endgame has only arisen 118 times out of 7,187,893 games so it is extremely rare. Luckily it has arisen in a World Championship game. That two bishops always win against a knight was one of the important discoveries made by Tablebase. Black can decide the game either by winning the knight or by playing for mate. This position arose in the seventeenth game in the World Championship match between Botvinnik – Tal in 1961. Tablebase informs us that it’s mate in 23 moves and Tal won quickly after 77...♗f1+ 78 ♔b6. Tablebase’s variation goes: 78 ♔b7 ♔d6 79 ♔b6 ♗e2 80 ♘a5 ♗c5+ 81 ♔b7 ♗d4 82 ♘b3 ♗c3 83 ♔b6 ♗c4 84 ♘a5 ♗d4+ 85 ♔b7 ♗e6 86 ♔a6 ♔c5! 87 ♘b7+ ♔b4 A beautiful king manoeuvre which tightens the mating net. 88 ♘d6 ♗c5 White loses the knight since it must cover the c8-square. 89 ♔b7 (89 ♘b7 is met by 89...♗c4 mate.) 89...♗xd6 90 ♔c6 ♗c5 91 ♔c7 ♗d5 92 ♔d8 ♗c6 93 ♔c7 ♔b5 94 ♔d8 ♔b6 95 ♔c8 ♗e7 96 ♔b8 ♗d7 97 ♔a8 ♗e6 98 ♔b8 ♗d6+ 99 ♔a8 ♗d5 mate. 78...♔d6 79 ♘a5 ♗c5+ 80 ♔b7 ♗e2 81 ♘b3 ♗e3 The bishop pair is extremely strong on a board emptied of pawns and makes it impossible for White’s knight to do anything. 82 ♘a5 82...♔c5 Tal plays the pragmatic move and plans to pick up the knight. Tablebase prefers 82...♗d4 83 ♘b3 ♗c3 84 ♔b6 ♗c4 85 ♘a5 ♗d4+ 86 ♔b7 ♗e6 87 ♔a6 ♔c5!. This king manoeuvre should be remembered! 83 ♔c7 ♗f4+ and here Botvinnik resigned... ...since the knight is lost after 84 ♔d8 (84 ♔b7 ♔b5 85 ♘b3 ♗e3 makes no difference.) 84...♔b6 85 ♘b3 ♗e3 followed by ♗d1 forcing the knight into the corner when ...♗d4 follows. This is a useful ending to practice with the computer. You play with the bishop pair and try to beat the computer. You will learn a lot about how to play with the bishop pair and the technique you learn can be used in other and completely different positions where the bishop pair is involved. It’s about learning bishop pair technique! GM Glenn Flear has written an excellent book Practical Endgame Play – Beyond the basics – the definitive guide to the endgames that really matter and it’s one of the few books available focusing on the very important endgames where you have more than one piece each but with a maximum of two. Flear has a name for such endings and calls them ‘NQE’ – for ‘Not Quite an Endgame’ (pronounced ‘nuckie’). According to his statistics 15-20 % of all games reach the following type of ending. This means that if you are seriously studying such endings you will have a substantial advantage over players who haven’t incorporated such study in their habit of chess training. 205 Flohr – Botvinnik Match 1933 White to move According to Botvinnik this endgame should end in a draw although White can expect some practical winning chances on a scale of 60 to 40. White has more space as well as a pair of bishops against a pair of knights. The main drawback of the two knights is that it will be harder to exchange one of the bishops but Black should be able to draw with active play. There are two other things going for Black. Firstly, the position is locked. In an open position the bishop pair would have been a decisive factor with pawns on both sides of the board. Secondly the pawn position is symmetrical. White has no passed pawn(s) and for the moment no quantitative pawn majority when the bishop pair might be a crushing factor. The game continued: 25 ♔f2 ♔e7 26 ♗e3 ♔d8 27 ♔e1 ♔c7 28 ♔d2 ♘c5 29 b4 ♘cd7 30 g3 ♘b6 31 ♔c2 31...♘bd7? After this passive move it will be more difficult for Black to play in an active manner and so White’s winning chances increase accordingly. Presumably Black’s only chance to hold the draw is by preventing White’s activity on the queenside with 31...♘a4! 32 ♗d1 (32 ♔b3 b5) 32...b5 33 ♗xa4 bxa4 34 ♔c3 ♘d7 35 b5 a5!. It’s not easy for White to do anything constructive on the queenside. If White retains his bishop pair and activates his king on the kingside, Black’s knight will manoeuvre on the queenside with counter chances. 32 a4! Obviously White seizes space on the queenside. 32...♘b6 33 a5 The establishment of a strong pawn clamp makes Black’s defence more diffcult. Note that the b7-pawn has been fixed and is a future target for the light-squared bishop. 33...♘bd7 34 ♗c1 ♔d8 35 ♗b2 ♘e8 36 ♔d2 ♘c7 Botvinnik wants to improve his game with 36...♔e7, followed by ...g7-g6 and ...f7-f5, instead of allowing White the luxury of controlling the whole board. 37 ♔e3 ♔e7 38 ♗f1 ♘b5 39 h4 ♘c7 40 ♗h3 ♘e8 Here the game was adjourned and this was Botvinnik’s sealed move. 41 f4 f6 42 ♗f5 It’s typical of the bishop to provoke weaknesses in the enemy pawn position. 42...g6 43 ♗h3 Now White can profit from the weakness by planning f4-f5 and fxg6 followed by the creation of a passed pawn with g3-g4 and h4h5. This is an important idea to remember. 43...h6 44 ♗c1 ♘g7 45 fxe5 dxe5 The other recaptures aren’t playable: 45...♘xe5? fails to 46 ♗c8, threatening the b7-pawn, while 45...fxe5? loses to 46 ♔f3 with the idea 46...h5 47 ♗g5+ ♔e8 48 ♗h6 and one of the knights is lost. Such situations are not uncommon when dealing with a pair of bishops with a lot of space. 46 ♔f3 h5 47 ♗e3 ♔d6 48 ♗h6 ♘e8 49 g4 hxg4+ 50 ♗xg4 ♘c7? 50...♔e7 was better, planning ...♘d6. After 51 ♗e3 ♘d6 52 h5 gxh5 53 ♗xh5 f5 Black has drawing chances thanks to the many pawns that have disappeared from the board. 51 ♗e3 ♘b5 52 ♔e2 ♘c7 53 ♔d3 Black is in zugzwang since the c5-square must be covered, his d7-knight is preventing ♗c8 and the c7-knight stops the manoeuvre ♗e6-f7. 53...f5 54 exf5 gxf5 54...♘f6 55 ♗h3 gxf5 gains a tempo but Black is lost anyway because of the strong distant passed pawn. 55 ♗xf5 ♘xd5 56 ♗d2 ♘7f6 57 ♔c4 ♔c6 58 ♗g6 b5+ 59 ♔d3 ♘e7 Here the game was adjourned for a second time. 60 ♗e4+ ♘ed5 61 ♗g5 ♘h5 62 ♗f3 ♘g3 63 ♗d2 ♔d6 64 ♗g4 ♘f6 65 ♗c8 ♔c6 66 ♗e1 e4+ 67 ♔d4 ♘gh5 68 ♗f5 ♔d6 69 ♗d2 Black resigned. A very instructive endgame showing the difficulties the defending knights had to contend with after passive play. Note that Flohr avoided variations where one of the bishops might have been exchanged since that would have reduced his winning chances. His main strategy was to keep the bishops away from the knights and exploit his space advantage for decisive breaks. Every pawn exchange made the position more and more open, resulting in two overpowering bishops. This endgame illuminated what Tarrasch once said: “The future belongs to he who has the bishops.” 206-277: Rook Endings Rook endings demand the most knowledge and therefore quite a lot of positions are presented since these occur in 50% of all endgames. Another reason is that sometimes the correct procedures in rook endings are almost of a mystical and unique nature. Even sensible moves can turn out to be mistakes, which is not usually the case in other endgames. Another important piece of information is that rook endings can often be saved despite the fact that a player is one or even two pawns down. Holding the draw in difficult situations was actually one of Alekhine’s secrets. When you are in trouble you try to reach a rook ending a pawn down according to the old adage: “All rook endings are drawn”. 206 Jiretorn – Engqvist Kungsbacka 1980 White to move When playing against a pawn, the best position for the rook is behind it. If White’s position is sufficiently advanced then it’s possible to draw. Here I have just checked my opponent, hoping he would step in front of the pawn. 66 ♔e8! The correct move, even though the king is cut off from its own pawn. However, the most important thing is not to stand in the way of the runner. 66...♖e4+ 67 ♔f8? 67 ♔f7 was necessary, repeating the position. Then it would not have been possible to force the king in front of the pawn. After 67...♔d7 68 g7 ♖f4+ White plays 69 ♔g6!. 67...♔d7! White missed or underestimated the strength of this move. 68 g7 ♖f4+ White is forced to place his king in front of the pawn and now Black wins. 69 ♔g8 ♔e7 70 ♔h7 ♖h4+ 71 ♔g6 ♖g4+ 72 ♔h7 ♔f7 73 ♔h8 White tries a stalemate trick which is duly avoided by 73...♖h4 mate. 207 Réti 1929 White to move Rook against pawn can also be tricky when the rook is in front of the pawn. The solution to Réti’s famous study is 1 ♖d2! Surprisingly the natural move 1 ♖d1?, keeping the rook away from threats by the king, doesn’t work here. Black draws after 1...d4 2 ♔d7 ♔d5!! when White is in an amusing form of zugzwang; 3 ♖d2 (3 ♔c7 is met by 3...♔c5! and 3 ♔e7 by 3...♔e5! renewing the zugzwang.) 3...♔e4 4 ♔c6 ♔e3 Black gains the necessary tempo by threatening the rook. 5 ♖d1 d3 6 ♔c5 ♔e2 followed by ...d2 with a draw. White has an alternative win with 1 ♖d3! which after 1...d4 2 ♖d1 leads to the same position as in the main variation. 1...d4 2 ♖d1! This is the key position. Now Black is in zugzwang. 2...♔d5 3 ♔d7 ♔c4 If 3...♔e4 then the king would choose the other way with 4 ♔c6. 4 ♔e6 d3 5 ♔e5 ♔c3 6 ♔e4 d2 7 ♔e3 White is just in time to win the pawn. 208 Kopaev 1958 Black to move When the rook is facing two far advanced pawns it’s important that the king is close enough to enable it to save the draw. The best place for the rook in this study by the famous study composer Nikolay Kopaev is 1...♖f3!. It’s necessary to prevent White from playing ♔g7 followed by the march of the f-pawn. If you are not familiar with this type of position your instinct might fail you and the mistake 1...♖g3? will appear on the board. The problem is that it allows 2 ♔g7!. Dvoretsky calls this idea “a change of the leader” which is an important idea to remember, both for the attacker and the defender. After 2...♔d4 3 f6 ♔e5 4 f7 ♖f3 5 f8♕ ♖xf8 6 ♔xf8 ♔f6 7 g7 White promotes. 2 g7 If White tries to change the leader with 2 ♔e6, preparing f5-f6, then Black replies 2...♖g3! controlling both pawns. This is a typical method to keep two pawns under control when they are not too far advanced. 2 ♔e5 doesn’t change much after 2...♖g3 3 f6 due to the nice intermediate check 3...♖g5+!. But not 3...♖xg6? immediately since 4 f7 leads to a win after 4...♖g5+ 5 ♔e4 ♖g4+ 6 ♔f3 ♖g1 7 ♔f2. 2...♖g3 3 ♔f7 ♔d4 4 f6 4 g8♕ ♖xg8 5 ♔xg8 ♔e5 and the pawn is captured next move. 4...♔e5 This drawn position is important to store in your memory. It’s called “the tail-hook” by Dvoretsky, since Black’s rook controls the g-pawn while the black king puts pressure on the f6-pawn. 209 Tarrasch – Janowski Ostend 1907 Black to move Here Black should put pressure on the leading pawn by 81...♖f1!. The rook effectively freezes both pawns in their tracks. 82 ♔d4! White keeps the opponent’s king out of c3 while preparing ♔e5 supporting the f-pawn. 83...♔b3 82...♖f5 doesn’t work on account of 83 ♔e4 ♖xg5 84 f7 ♖g4+ 85 ♔f3 ♖g1 86 ♔f2 leading to the ending queen against rook. 83 ♔e5 ♔c4 84 g6 ♖e1+ 85 ♔d6 ♖g1 This is the toughest defence. The game continued with the useless checks 85...♖d1+ 86 ♔e6 ♖e1+ and after 87 ♔f7 Janowski resigned. 86 g7 ♔d4 210 White to move 87 ♔c6!! The point of this move is that it threatens f6-f7 since White’s king is ready to go to b5 after a rook check on g6. a) 87 f7? immediately fails to 87...♖g6+ and White is forced to enter the seventh rank when Black captures the g7-pawn and pins the f7-pawn. b) 87 ♔e6? is answered by 87...♔e4 88 ♔f7 ♔e5 and the pressure on both pawns secures Black the draw (“the tail-hook”). c) 87 ♔e7? ♔e5 is likewise a draw. 87...♔c4! The best defence since after 87...♖g6?, 88 ♔b5 followed by f7 is decisive. 88 ♔d7! White’s king simply walks down to the eighth rank followed by f6-f7 with a win, exploiting the lack of pressure on the f6-pawn. Compare this variation with variation (c). 88...♔d5 89 ♔e8! ♔e6 90 f7 ♖a1 Black tries a last shot. But alas after 91 f8♘+! ♔f6 92 g8♕ ♖a8+ 93 ♔d7 White wins. It’s a pity Janowski didn’t defend in a more accurate way because it would have been enjoyable to watch Tarrasch’s execution of this beautiful variation. 211 Moravec 1924 White to move This position would be a clear loss for White if it weren’t for the awkward position of the black king. After 1 ♖a1+ ♔b8 2 ♖b1+ ♔c8 3 ♖a1 ♔d8 4 ♔d6 ♔e8 5 ♔e6 ♔f8 6 ♔f6 ♔g8 7 ♖a8+ ♔h7 8 ♖a7+ ♔h6 9 ♖a8 ♔h5 10 ♔f5 ♔h4 11 ♔f4 the black king cannot escape the perpetual mating threats and has to go back with 11...♔h5 when the position is repeated after 12 ♔f5. This idea only works with the pawns on the rook and knight files, otherwise Black would be able to promote with check. For example if Black had the h2-pawn on f2 instead, there is ...f1♕+ in the variation where White plays the king to f6. 212 Kamsky – Bacrot Sofia 2006 Black to move This ending often arises from the ending rook versus pawn. This is what happened in the game. Black was unable to queen the epawn due to mate and promoted to a knight with check. White has just played 75 ♔g3. It’s normally easy to hold the draw for the defender, the main reason being that it’s difficult for the attacker to place his king in front opposition because the knight often controls this square. Bacrot played 75...♘d3 76 ♖d2 ♘e1 This is the simplest way to draw, keeping the knight next to the king while controlling the opposition square f3. 77 ♖f2+ ♔g1 78 ♖f8 Black’s king is cut off but it doesn’t mean anything. Most important is to keep the knight close to the king. 78...♘g2 The seemingly sensible 78...♘d3? is actually a trap after 79 ♔f3!. It’s impossible for the knight to reunite with the king since 79...♘e1+ (79...♔f1 is the move you want to play but it fails to 80 ♔e3+ and this is the main reason 78...♘d3 is a mistake.) 80 ♔e2 ♘g2 is met by 81 ♖e8! ♘h4 82 ♖e4 and White wins the knight in a few moves. 78...♘c2! is the easiest way to draw, planning to go back with 79...♘e1. For example 79 ♔f3 (or 79 ♖e8 ♔f1) 79...♔f1. 79 ♔f3 ♔f1? 79...♘h4+ was the only move to draw. After 80 ♔e3 ♔h2 81 ♖g8 ♔h3 the typical drawing position has been reached. 80 ♔g3+? 80 ♖f7! or 80 ♖f6! wins. 80...♔g1 81 ♔f3 ♔f1? 81...♘h4+!. 82 ♖f7! ♘e1+ 83 ♔e3+ ♔g1 84 ♔e2 ♘g2 85 ♖h7 ♘f4+ 86 ♔f3 ♘d3 87 ♖h4 ♘e5+ Reunion with the king by 87...♘e1+ doesn’t work after 88 ♔e2 ♘g2 89 ♖g4 ♔h2 90 ♔f2 and the knight is lost. 88 ♔e2? Correct was 88 ♔e3 or 88 ♔e4. Now it’s a draw. Presumably the players were in time pressure, judging from the high proportion of errors in this theoretically drawn ending. 88...♔g2 89 ♖e4 89...♘f7? 89...♘d7 is a draw after 90 ♖e7 ♘f6 91 ♖g7+ ♔h3 92 ♔f3 ♔h4 and the knight is ready to reunite with the king after 93...♘h5. The main thing to remember is to place the knight next to the king and control the opposition square. 90 ♖e7 ♘d6 91 ♖g7+ ♔h3 92 ♔f3 ♔h4 93 ♔f4 ♔h5 94 ♖e7 More precise was 94 ♖c7!, for example after 94...♘e8 95 ♖e7 ♘f6 96 ♔f5 ♘g4 97 ♖h7+ ♘h6+ 98 ♔f6 a typical zugzwang has arisen. 94...♘c4 95 ♖e6 ♘d2 96 ♖c6 ♘b3 97 ♔e3 ♔g4 98 ♖c4+ ♔g3 99 ♖c3 ♘a5 100 ♔e4+ ♔f2 101 ♔d5 ♘b7 102 ♖b3 ♘d8 103 ♖b8 Black resigned. Don’t forget to practice this endgame because it’s pretty common in competitive play. 213 Em. Lasker – Ed. Lasker New York 1924 White to move Edward Lasker thought he was going to win with rook and pawn versus a knight, when his dream was shattered by Emanuel Lasker’s continuation 93 ♘b2!. He had only reckoned on 93 ♔a3 ♔e4 94 ♘c5+? ♔d4 95 ♘xb3+, whereupon 95...♔c4 would have won the knight. However, as Emanuel Lasker proves in the game, there is no hurry to win the pawn. Play continued: 93...♔e4 94 ♘a4 ♔d4 95 ♘b2 ♖f3 96 ♘a4 ♖e3 White is going back and forth between a4 and b2, exploiting the fact that the black king cannot cross the third rank. 97 ♘b2 ♔e4 98 ♘a4 ♔f3 99 ♔a3! White changes plan at just the right moment since Black’s king is planning to come behind the pawn. However, White simply plays ♔b2 followed by ♘c5-♘xb3. Note that 99 ♘b2? is a mistake due to 99...♔e2 100 ♘c4 (100 ♔a3 ♔d2 101 ♘c4+ ♔c1 102 ♘b2 ♖g3 and White is placed in zugzwang.) 100...♖g3 101 ♔a3 ♖c3! 102 ♘b6 (102 ♘b2 ♔d2) 102...♔d3 103 ♔b2 ♖c6! 104 ♘a4 ♔c4 and Black eventually wins. 99...♔e4 99...♔e2 100 ♔b2 ♔d2 101 ♘c5 and the b3-pawn falls with a drawn result. 100 ♔b4 ♔d4 101 ♘b2 ♖h3 102 ♘a4 ♔d3 103 ♔xb3 ♔d4+ The game was in its fourteenth hour and Edward Lasker had to reconcile himself to a draw. 214 Vidmar – Alekhine San Remo 1930 White to move As a rule this type of ending is a draw but the defence can prove to be difficult in practical chess. Vidmar played 39 h4 which is a good move, taking space and activating the white position. 39...♔e7 40 ♘e4 h6 41 ♘f2? White plays too passively and that is dangerous in this type of ending. The best chance to achieve a draw is by active play and striving to pressurise the black pawns. 41 ♔h3!, preparing g2-g4, makes sense. a) 41...♖a3 makes it possible for White to play 42 ♔g4, for example 42...♔e6 (42...♖a2 43 g3) 43 ♔f4 ♔d5 44 h5. Now the g7pawn is fixed. The next step is to manoeuvre the knight and put pressure on the g7-pawn, thereby reducing one of Black’s pieces to passivity. Then 44...♔d4 45 ♘d6 ♔d3 46 ♘f5 ♖a4+ 47 ♔g3 ♖a5 48 ♔f4 ♖a7 49 g4 ♔e2 50 ♔g3 leads to a draw according to Leikin. b) 41...♔e6 42 g4! ♔e5 43 ♔g3 g6 Black doesn’t allow the g7pawn to be fixed by h4-h5. 44 ♘f2 ♖a3 45 ♘h3 ♖a4 Black shouldn’t allow White’s knight to reach f4. 46 ♔f2! White has constructed a kind of fortress and it’s impossible for Black to stop ♔g3 and then ♘f4 or ♔g3 followed by ♔f2. 46...h5 47 gxh5 gxh5 (47...♖xh4 48 hxg6 is a draw.) 48 ♔g3 Black cannot break through and must think about how to defend the pawn weakness on h5. 41...♔e6 42 ♘d3 ♔f5 43 ♘f4 ♖a4 44 ♘d3 ♖c4 45 ♘f2 ♖c6 46 ♘h3 ♔e5 47 h5? White’s knight is too far away from the f5-square and the gpawn is not placed on g4 so this move is not in harmony with the rest of the position. It’s better to play along the lines 47 ♘f4 ♖c2 48 ♘h3 ♖d2 49 ♘f4 ♖a2 50 ♘h3 ♔d4 51 ♘f4 ♔e3 52 ♘e6 ♖a7 53 ♘f4 ♖a6! 54 ♘h3 ♔e2 55 ♘f4+ ♔f1 56 ♘d5 and hope that Black will not be able to break through. However, this is the kind of play you want to avoid as a defender in view of Black’s very active king. 47...♖c2! 48 ♘f4 48...♖d2? 48...♔f5! was correct. 49 ♘h3? Once again overly passive play by Vidmar. Miles’ recommendation 49 ♘g6+! was better and after 49...♔d4 to continue 50 ♘h4, planning ♘f5 as well as ♔f4 with excellent drawing chances. 49...♔d4 50 ♘f4 ♔e3 51 ♘e6 Waiting play with 51 ♘h3 doesn’t work on account of 51...♖b2 52 ♘f4 ♖b5! 53 ♘e6 ♖e5! and the g7-pawn is taboo. 51...♖d5! Such a finesse with the rook is typical and one main reason why these endings are so tricky to handle. It’s vital for White to exert more effective pressure on the opponent’s pawns as already shown in some of the above mentioned variations. 52 f4 52 ♔h4 ♖e5! 53 ♘xg7 ♖g5 54 ♘e6 ♖xg2 leads to a loss for White. The f-pawn is doomed. 52...♖f5! 53 ♔g4 ♖f6! 54 f5 ♖f7 It’s a matter of taste whether Black chooses this or 54...♔e4 55 ♘xg7 ♖f7 56 ♘e6 ♖xf5. 55 g3 55 ♘d8 ♖d7 56 ♘e6 ♔f2 and White’s g-pawn is lost. 53...♔e4 54 ♘c5+ ♔d4 55 ♘b3+ ♔e5 and Black resigned. The moral of this ending is, if you are saddled with the knight, play more actively than Vidmar did. 215 Vedder – Engqvist Novi Sad 2016 White to move Because of stalemate issues, rook against bishop is an easy draw if the defender’s king is in a corner of the opposite colour to the bishop. In this position White only needs to think about placing the bishop at a sufficient distance from the black king. 107 ♗g6 One of the traps to be aware of is 107 ♗d3?? ♖a2+ 108 ♔b1 ♖d2 and the defender has no saving check with the bishop due to being too close the opponent’s king. The defender has to be careful not to fall for this simple trick. The defensive rule is to place the bishop far away so that it’s possible to check the attacker’s king along the appropriate diagonal when it’s in front of the black king. 107...♖a2+ 107...♖e1+ 108 ♗b1 ♔a3 is stalemate. It’s impossible to create conditions for a zugzwang when the king is in the corner and the bishop in front of it, but if White has a bishop on c1 and the king on b1, the available a1-square means the position is lost. 108 ♔b1 ♖g2 109 ♗f7+ The check on the classical diagonal is vital to save the draw. 109...♔c3 110 ♔a1 ♔c2 111 ♗e6 Of course not 111 ♗a2?? ♖g1+ and Black wins. 111...♖f2 112 ♗g8 ♖f8 113 ♗d5 ♖f4 114 ♗e6 ♖f1+ 115 ♔a2 ♖f6 116 ♗b3+ The saving check once again. 116...♔c3 117 ♗g8 ♖a6+ 118 ♔b1 ♖a8 119 ♗e6 ♖b8+ 120 ♔a1 ♖e8 121 ♗f7 ♖e1+ 122 ♔a2 ♖g1 123 ♗e6 ♖g6 Here my opponent claimed a draw according to the 50 move rule. 216 Horwitz/Kling 1851 White to move A rook and a pawn on the sixth rank can sometimes be difficult to win, since it’s not easy for the king to participate effectively. On the sixth rank the pawn stands in the way of its own king. However, this study by Horwitz and Kling is a win. 1 g7! White’s g6-pawn is in the way of the king so it’s best to sacrifice it. 1...♔h7! 1...♗xg7 leads to a quicker win after 2 ♔g6. If Black tries to save his bishop it leads to mate after 2...♗e5 (or 2...♔h8 3 ♖b8+) 3 ♖e7 ♗d6 4 ♖e8+ ♗f8 5 ♖d8 ♔h8 6 ♖xf8 mate. 2 ♖f7! The key idea is to trap the king so it cannot escape via the f-file. 2...♗d4 3 g8♕+! ♔xg8 4 ♔g6 Black cannot move the king to f8 and now it’s only a matter of setting up a decisive double threat against the king and the bishop. 4...♗g1 Black’s bishop moves as far away as possible from the rook’s influence. 5 ♖f1 ♗h2 6 ♖f2 Another win is 6 ♖h1 ♗g3 (6...♗f4 7 ♖a1) 7 ♖h3 ♗f4 8 ♖a3 ♔f8 9 ♖f3. 6...♗g3 7 ♖g2 ♗d6 7...♗e5 8 ♖e2. 8 ♖d2 ♗e7 9 ♖c2 Black cannot prevent the decisive ♖c8+ next move. 217 Szabó – Botvinnik Budapest 1952 White to move 66 ♖b4 66 f7 is met by 67...♔g7! and the position is a draw since Black captures the pawn next move. An immediate 66...♗xf7? loses to 67 ♔f6. 66...♗a2 The only move to hold the fortress. 67 ♔f5 ♗d5 The only move again. 67...♔f7? loses to 68 ♖b7+ ♔f8 69 ♔g6 since Black has no check with his bishop. 68 ♔g6 ♗f7+ 69 ♔g5 ♗d5 70 ♖h4 ♗b3 71 ♖h8+ ♔f7 72 ♖h7+ ♔f8 73 f7 ♔e7! The main defensive idea, which prevents White’s king from going to f6. 74 ♔g6 ♗c4 If Black retains the pressure on the f7-pawn it’s a draw, so any move on the a2-e6 diagonal is fine. Of course not 75...♗c2+? 76 ♔g7. 75 ♖g7 ♗d5 76 f8♕+ ♔xf8 77 ♔f6 ♔e8 78 ♖e7+ ♔d8! Note that White wins a similar position with a central pawn or a knight pawn, even if it’s placed on the sixth rank. 218 Dvoretsky 2003 White to move Black has set up an impregnable fortress. It’s important to place the h-pawn on h5 before White gets the chance to place his g-pawn on g4. If Black has placed the pawn on h5 it doesn’t matter if White’s g-pawn is placed on g5. Dvoretsky gives the following variations: 1 h4!? When the pawn is on g3 White can try to manoeuvre the king to g5 but Black will stop that from happening by playing ...♗f6 when White’s king is on f4. An illuminating variation is 1 ♔d5 ♗f6 2 ♔e4 ♗c3 3 ♔f4 ♗f6 4 g4!? hxg4 5 ♔xg4 ♗e5 6 h3 ♗f6 7 ♔f4 ♔f8 8 ♔e4 ♔g8 9 ♔d5 ♗a1 10 ♔e6 ♗c3 11 ♖c7 ♗b2 12 h4 ♗d4 13 ♖c4 13...♗e3!. This is the only move to hold the draw. (Murey – Ftáčnik, New York 1987 continued with the fatal 13...♗b2? when White went on to win after 14 ♖g4 ♔h7 15 ♔f7 This trap is important to remember for both players. 13...♗f2? also loses after 14 ♔f6 ♔h7 15 ♖g4 ♔h6 16 ♖xg6+ ♔h5 17 ♖g2.) 14 ♔f6 (14 ♖e4 ♗d2 15 ♖e2 ♗c3 and Black holds.) 14...♔h7 15 ♖g4 ♔h6 16 ♖xg6+ ♔h5 17 ♖g3 ♗b6!! 18 ♖h3 ♔g4 19 ♖h1 ♗d8+ and Black achieves a draw. Back to the main variation: 1...♗a1! The bishop must hide from the rook. 1...♗d4? 2 ♖b4 gains an important tempo. Compare with the main variation. 2...♗c3 3 ♖c4 ♗b2 (3...♗e1 4 ♔f6 ♔h7 5 ♖c6! ♗xg3 6 ♔g5 and Black is in zugzwang.) 4 g4 hxg4 5 ♖xg4 ♔h7 6 ♔f7 ♔h6 7 ♖xg6+ ♔h5 8 ♖g2! ♗c3 9 ♖h2 ♗d4 (9...♗e1 10 ♔f6 ♗xh4+ 11 ♔f5) 10 ♔e6 ♗g1 11 ♖h1 ♗f2 12 ♔f5 and White has a theoretical win. 2 ♖b4 ♔g7! 3 g4 hxg4 4 ♖xg4 ♔h6 5 ♖g5 5 ♔f7 ♔h5 is an easy draw. 5...♗d4 6 ♔f7 ♗f2 7 ♖xg6+ ♔h5 and it’s a draw. A very instructive endgame since you can also learn from it related positions with h-pawn versus g-pawn. This is the advantage of studying endgames with several pawns. 219 Guretzky/Cornitz 1863 White to move The key to winning this ending is to sacrifice the h-pawn at the right moment, while preventing Black’s king from escaping to one of the safe corners on a8 or h1. 1 ♔g4 Step one is to improve the king as much as possible. 1...♔g7 2 ♔f5 ♔h6 3 ♖a3 Step two is to cut off Black’s king and force it to the seventh rank. 3...♔h7 3...♗c1 4 ♖a6+ makes no difference since 4...♔xh5 is unplayable due to 5 ♖a1. 4 ♖a6 ♔g7 5 h6+! When White’s pieces are optimally placed the conditions are ripe for this pawn sacrifice. 5...♔h7 5...♗xh6 6 ♖g6+ ♔h7 7 ♔f6 followed by ♔f7 is winning since the black king is trapped in the wrong corner of the board. 6 ♖g6! Step three is to transfer the king to h5 or f7. 6...♗f4 6...♗c3 prevents the rook check but not 7 ♔g5 followed by ♔h5. 7 ♖g7+! White forces the king to take the h6-pawn. 7...♔xh6 8 ♖g6+ ♔h7 8...♔h5 9 ♖g8. 9 ♔f6 followed by ♔f7 with an elementary win if you are familiar with the famous study by Kling and Horwitz from 1851, position 216. 220 Studenetsky 1939 White to move This position shows how strong the control of the absolute seventh rank can be. White to play wins after: 1 ♖g8+ ♔d7 2 ♖g7+ ♔e8 2...♔d8 3 ♔d6 leads to mate or wins the rook. 3 ♖h7!! An unbelievable move facing Black with zugzwang. 3...c4 4 ♔d6! White threatens ♖h8+ so Black must react, but alas all moves are losing, for example 4...♖d8+ 5 ♔e6. Providing you don’t forget this position you’ll always think twice to see if there is any magic involved when controlling the seventh rank. This is something you should do whether it’s an endgame or a middlegame since the rook’s normal value has increased by quite a considerable extent. 221 Philidor 1777 Black to move Of all the endings you’ll play in your life approximately 50 per cent of them will be rook endings. But if you only know one important position then it should be the Philidor position! Black to move draws with 1...♖h6! It’s important to cut off the enemy king so it’s unable to reach the sixth rank. 2 e5 ♖g6 Black keeps his rook on the sixth rank until White advances his pawn there. 3 e6 White is threatening to penetrate decisively with the king, so Black must play offensively with the rook. 3...♖g1! 4 ♔d6 ♖d1+ This is the saving idea. White cannot hinder the checks from the rear and is unable to create any danger for the defender. 5 ♔e5 ♖e1+ 6 ♔f6 ♖f1+ It’s a draw since White cannot drive away the black king from either e8 or e7. This is a simple and good illustration of Lasker’s saying that “He who can play with a rook can play chess.” 222 Lucena 1497 White to move The second most important rook ending to know by heart is the Lucena position from the 15th century. White wins with 1 ♖f4!. White needs to build a bridge for the king to be able to win. It’s too early for 1 ♔e7?! ♖e2+ 2 ♔d6 ♖d2+ 3 ♔e6 ♖e2+ 4 ♔f5 ♖d2 which leads nowhere. White must start all over again and build a bridge after 5 ♔e6 ♖e2+ 6 ♔d6 ♖d2+ 7 ♔e7 ♖e2+ 8 ♔d8 ♖c2 and now 9 ♖f4!. 1 ♖f5! also wins but Black can cause some irritation with his king. Remember that the critical squares are f4, f5 and f8 and that Black cannot control all these squares with the king! 1...♔g6 2 ♖f4 (Or 2 ♖f8 ♔g7 3 ♖f4! building a bridge.) 2...♔g5 3 ♖f8! ♔g6 4 ♔e7 ♖e2+ 5 ♔d6 ♖d2+ 6 ♔c6 ♖c2+ 7 ♔b5. In this last variation White didn’t need a bridge. 1...♖c1 2 ♔e7 The right time to step out with the king. 2...♖e1+ 3 ♔d6 ♖d1+ 4 ♔e6 4...♖e1+ 4...♖d2 looks more clever but after 5 ♖f5! and ♖d5 White still builds a bridge – but on the fifth rank. 5 ♔d5 ♖d1+ 6 ♖d4 White’s bridge (or umbrella) is completed and there are no more annoying checks (or pouring rain!). 223 Philidor 1777 Black to move Black threatens to win with 1 ♖h8+ ♔d7 2 e6+ so Black must put pressure on the e-pawn at once. 1...♖e1! 2 ♔e6 2 e6 is met by 2...♖f1+ in accordance with the previous Philidor’s position, while 2 ♖h8+ is useless after 2...♔d7 since White cannot advance his e-pawn to e6. This is the main point of placing the rook on the e-file. 2...♔f8! The rule is to step to the short side with the king and keep the long side for the rook. In this position 2...♔d8 draws as well, since after 3 ♖h8+ ♔c7 4 ♖e8 ♖h1 5 ♔f7 ♖h7+ followed by 6...♔d7 the long side is long enough for the rook to make disturbing checks. However, when the stronger side has the bishop pawn it’s vital to step with the king to the short side of the pawn so the long side is available for timely annoying checks, otherwise the position is lost since it will transpose to a Lucena position. 3 ♖h8+ ♔g7 4 ♖e8 White threatens to play the king to d7 so this is the signal to transfer the rook to the long side and upset White’s plan of creating a Lucena position. After 4 ♖a8 Black maintains the pressure on the f-pawn with 4...♖e2! since 5 ♔d6 is answered by 5...♔f7. 4...♖a1! 5 ♖d8 5 ♔d7 ♖a7+ 6 ♔c6 ♖a6+ leads to an array of checks followed by ...♔f7 threatening the rook at the right moment. 5...♖e1! 6 ♔d6 ♔f7 There is no way for White to win and he can only repeat what has gone before with 7 ♖d7+ ♔e8 8 ♖a7 ♖e2 9 ♔e6 etc. 224 Vancura 1924 White to move The idea of using the long side for the rook is also the main method to draw this position by Vancura, which could be regarded as the third most important endgame after Philidor and Lucena. After 1 ♔e5 Black must maintain the pressure against the a6pawn so White cannot liberate his rook. Any move along the sixth rank is suitable. When White approaches the enemy rook, Black goes back with the rook to f6 followed by checks along the f-file. Play can go like this: 1...♖b6 2 ♔d5 ♖f6! The simplest move. Black’s king is best placed on g7 or h7 and doesn’t need to be moved. Beware of the tactical trap 2...♔f7?? 3 a7 ♖a6 4 ♖h8! ♖xa7 5 ♖h7+ and White wins the rook. 3 ♔c5 3...♖f5+ The rook should be placed on the f-file when delivering checks. 3...♖h6 loses to 4 ♔b5 ♖h5+ 5 ♔b6 ♖h6+ 6 ♔b7 since there is no available check on h7 because the black king is standing in the way. It’s too late for 6...♖f6 as White replies 7 ♖c8 and the liberation of the rook makes it easy for White to promote the pawn. 4 ♔b6 After 4 ♔d4 Black calmly plays 4...♖f6! maintaining the pressure on the a6-pawn. 4...♖f4+? loses after 5 ♔e5 ♖f6 (5...♖a4 loses as well after 6 ♔d5.) 6 ♖g8+. 4...♖f6+ Now it’s necessary to check the king since White would otherwise liberate his rook. 5 ♔a7 ♖f7+ 6 ♔b8 ♖f8+ 7 ♔b7 ♖f7+ 8 ♔c6 ♖f6+ 9 ♔d5 ♖b6 Again the simplest. 10 a7 ♖a6 11 ♔c5 ♖a1 12 ♔b6 ♖b1+ 13 ♔c7 ♖a1 White’s king cannot escape the checks from the rook when defending the pawn so the position is a dead draw. 225 Karstedt 1909 White to move When the king is in front of a rook pawn it’s always a win if White can reach the b8-square before the enemy king is placed on c7. 1 ♖c8 ♔d6! The toughest defence. 1...♔d7 loses after the prosaic 2 ♖b8 ♖h1 3 ♔b7 ♖b1+ 4 ♔a6 ♖a1+ 5 ♔b6 ♖b1+ 6 ♔c5. 2 ♖b8 ♖h1 3 ♔b7 ♖b1+ 4 ♔c8! 4 ♔a6 is a step in the wrong direction after 4...♖a1+ 5 ♔b6 ♖b1+. This is the main point of Black’s first move. 4...♖c1+ 5 ♔d8 ♖h1 6 ♖b6+! 6 ♔e8? doesn’t work on account of 6...♖h8+ 7 ♔f7 ♖h7+ and the pawn falls. 6...♔c5 6...♔e5 7 ♖a6 wins easily. 7 ♖c6+! This is an important idea to remember since it’s the only way to win this position. 7 ♖a6? leads to a draw after 7...♖h8+ 8 ♔d7 ♖h7+ 9 ♔e8 (Beware of 9 ♔e6?? ♖h6+ when Black wins!) 9...♖h8+ 10 ♔f7 ♖a8 11 ♔e7 ♔b5 12 ♖a1 ♔b6 13 ♔d6 ♖xa7 14 ♖b1+ ♔a5! (Of course not 14...♔a6?? 15 ♔c6.) 15.♔c5 ♖c7+ and it’s a draw. 7...♔b5 8 ♖c8 ♖h8+ 9 ♔c7 ♖h7+ 10 ♔b8 and White queens next move. 226 Troitzky 1896 White to move When the rook is in front of a rook pawn on the seventh rank, everything depends on the position of the black king. A win seems impossible but it can be achieved thanks to a tactical resource. 1 ♔f4 The king is on its way to c5 so other routes were possible as well, for example 1 ♔f5. The main thing is to lure the black king to c3 since it must follow the white king there to avoid a decisive rook check followed by a8♕. 1...♔f2 2 ♔e4 ♔e2 3 ♔d4 ♔d2 4 ♔c5 ♔c3 Rook checks are useless since after 4...♖c1+ 5 ♔b4 ♖b1+ 6 ♔a3 ♖a1+ 7 ♔b2 White is threatening not only the rook but also ♖d8+ followed by promotion. 5 ♖c8!! This is the whole point. 5...♖xa7 6 ♔b6+ White wins the rook after the discovered check. A clever tactical idea by the great endgame composer Troitzky for you to remember! 227 Grigoriev 1937 White to move In this position Black’s king is cut off from the b-pawn and the rook must do all the hard work to achieve the desired draw. 1 ♔a4 1 ♖c4 ♖a8 prevents White’s king from entering the a4-square. If 2 b5 ♔d7! 3 ♔b4 ♖c8 and Black draws. Note that 1...♔d7?, with the idea of ...♖c8, loses after 2 ♔a4 ♖a8+ (2...♖c8 3 ♖xc8 ♔xc8 4 ♔a5! and White gains the opposition.) 3 ♔b5 ♔d6 (3...♖b8+ 4 ♔a6 shows the point of having the rook on c4.) 4 ♔b6 ♔d5 (4...♖b8+ 5 ♔a7 ♖b5 6 ♔a6 and White wins by pushing the bpawn.) 5.♖c5+ ♔d4 (Or 5...♔d6 6 ♖c6+ ♔d7 7 b5) 6 ♖c6 ♖b8+ 7 ♔a5 ♖a8+ 8 ♖a6 ♖c8 9 b5 ♔c5 10 ♖a7 ♖h8 11 ♖c7+ ♔d6 12 b6 and White wins. 1...♖a8+ 2 ♔b5 ♖b8+ 3 ♔a5 ♖a8+ 4 ♔b6 ♖b8+ Black has enough space on the band a-files to force the white king back to b3. White cannot make any headway. Note that if the black rook had been placed on b7 this method wouldn’t have worked since the files would have been too short. 228 Grigoriev 1937 White to move If the opponent’s king is cut off by at least three files (the c-, dand efiles), counting from the b4-pawn, it’s always a win. It doesn’t matter where on the f-file the king is placed. The key to winning this ending is to have the d6-square available (two squares diagonally from the b4-pawn) for the king. 1 ♔c4 ♖c8+ 2 ♔d5 ♖b8 3 ♔c5 ♖c8+ 4 ♔d6! The ideal square for the king. 4...♖b8 5 ♖b1 ♔e8 6 ♔c7 ♖b5 7 ♔c6 ♖h5 7...♖b8 8 b5 ♔d8 9 ♖d1+ ♔c8 10 b6 and White wins due to the poor state of the black rook. 8 ♖d1! ♖h6+ 9 ♖d6 and White wins by pushing the b-pawn. 229 Lasker 1890 White to move White has the more active king and this factor decides this endgame. White plays 1 ♔b8 forcing 1...♖b2+. After 2 ♔a8 ♖c2 White creates more room for his king by 3 ♖f6+ ♔a5 and then repeats the same procedure. After 4 ♔b7 ♖b2+ 5 ♔a7 ♖c2 6 ♖f5+ ♔a4 7 ♔b7 ♖b2+ 8 ♔a6 ♖c2 9 ♖f4+ ♔a3 10 ♔b6 ♖b2+ (Otherwise ♖xf2 follows.) 11 ♔a5 ♖c2 12 ♖f3+ ♔a2 White plays 13 ♖xf2! exploiting the pin along the second rank. A beautiful and instructive study. 230 Gothenburg – Stockholm Black to move If Black has an active rook there are chances of holding the draw even when two pawns down. 1...♖a4! The idea is that White cannot activate his rook and king without losing a pawn and then Black holds the draw, since the remaining one is a rook pawn. 2 ♖f3+ ♔g6 3 ♔f2 White chooses to sacrifice the pawn on the kingside and exploit the fact that Black’s king is cut off. However, the position is still a draw. 3...♖xh4 4 ♔e2 ♖a4 5 ♔d2 ♖a6 6 ♔c2 ♖f6! The saving idea. The pawn ending is an easy draw after the exchange. 7 ♖d3 ♔f7 8 ♔b3 ♔e7 9 a4 ♖d6! and Black holds the draw without any problems. 231 Kasparov – Short World Championship match, London 1993 Black to move Rooks and two isolated pawns normally win by getting one pawn to the fifth rank and abandoning the other so that after its capture the defending king is cut off. In the ninth game of their World Championship match Short missed a draw against Kasparov. Here Short should have played: 46...♖c5!. Instead he played the fatal 46...♔e6?? and resigned after 47 ♔e3 ♔d6 48 ♔d4 ♔d7 49 ♔c4 ♔c6 50 ♔b4 ♖e5 51 ♖c1+ ♔b6 52 ♖c4. The king is cut off and White wins with the e-pawn. 47 ♖a3 White prevents the check on the third rank. 47 a5 doesn’t work either on account of 47...♖c3+ 48 ♔g4 ♔xe4 49 a6 ♖c8 50 a7 ♖a8 51 ♖a5 ♔d4 52 ♔f5 ♔c4 53 ♔e6 ♔b4 54 ♖a1 ♔c5!. The only move, shouldering the white king. 55 ♔d7 ♔b6 56 ♖b1+ ♔c5! This is the main defensive point. (56...♔xa7?? loses to 57 ♔c7.) 57 ♖b7 ♖h8! and White cannot improve his position any further. 47...♖c4! Such a move is easy to overlook if you are not familiar with the idea. It shows the defender’s rook’s ability to work both vertically and horizontally while making use of a check to gain a decisive tempo. 48 a5 ♖xe4 49 a6 49...♖f4+! Without this check Black is lost. 49...♖h4?? 50 ♖a5+ followed by a7 wins. 50 ♔g3 ♖f8 51 a7 ♖a8 The position is a draw. A likely end is 52 ♖a6 ♔d5 53 ♔f4 ♔c5 54 ♔e5 ♔b5 55 ♖a1 ♔b6 56 ♔d6 ♖xa7 57 ♖b1+ ♔a5! But not 57...♔a6?? 58 ♔c6 and Black must lose the rook to avoid mate in one. 58 ♔c5 ♖c7+ An instructive example showing the principal defensive capabilities of the rook. 232 Van Wely – Nguyen Olympiad, Baku 2016 White to move Here the pawns are more separated and the centre pawn is further down the board. White can nevertheless draw by exploiting the fact that the d3-pawn can be captured while the king is inside the square of the h-pawn. 92 ♖h1! White takes the roundabout route to capturing the d3-pawn while in the process controlling the h-pawn. The instant conquest of the pawn leads to a quick loss after 92 ♖c3?? h4 93 ♖xd3 h3. 92...♔f6 93 ♖h3 ♔g5 94 ♖xd3 Clever play by White has brought him the d3-pawn, while Black’s remaining pawn will be rendered harmless. 94...♖f5 The transition to the pawn ending 94...♖xd3+ 95 ♔xd3 ♔f4 96 ♔e2 ♔g3 97 ♔f1 leads to a theoretical draw and 94...♖e5 is answered by 95 ♖e3. 95 ♔e2 ♔g4 95...h4 96 ♖f3! is a well-known method to disarm the enemy rook’s control of the f-file. 96 ♖d8 h4 97 ♖g8+ ♔h3 98 ♖g7 ♔h2 99 ♖g4 h3 100 ♖g8 ♖a5 101 ♔f1 ♖a1+ 102 ♔f2 ♖a2+ 103 ♔f1 ♖g2 104 ♖f8 ♖g7 105 ♖f2+ 105...♔g3 105...♔h1 106 ♖f8 h2 107 ♖f6 ♖g1+ 108 ♔f2 ♖g2+ 109 ♔f1 and Black is unable to win since the king cannot get out of the corner. 106 ♔g1 ♖a7 107 ♖g2+! A typical drawing method to remember. 107 ♖f1 is a theoretical draw as well. 107...hxg2 It’s stalemate. 233 Gligorić – Smyslov Moscow 1947 White to move The ending with rook pawn and bishop pawn occurs relatively often and is very important to master. Although it’s a theoretical draw the attacker wins from time to time. One of the most famous examples is the game Carlsen – Aronian, FIDE Knockout, Tripoli 2004, where Carlsen lost. This proves that this endgame can be tricky with only 30 seconds per move. This ending frequently leads to endings where the h-pawn is used as a decoy in attempts to queen the f-pawn. In this famous rook ending between Gligorić and Smyslov the black rook works well on the rank by temporarily cutting off the king. 83 ♖g6+ White wants the king to declare its intentions. 83 f5 blocks the strong rook on the fifth rank but after 85...♖b1! the rook will nevertheless be active behind the enemy king and pawns. 83...♔f7! The basic rule for the king is that it should be placed on f7 until life becomes too dangerous and it runs the risk of being driven to the last rank. If that’s the case the king will be placed on g7. Later it can go to h6, blockading the h-pawn. 83...♔h7 doesn’t lose but makes the defence trickier. 84 ♖g5 The black rook is annoying for the white king so White forces it to change direction. 84...♖b1! This is an excellent square since, if necessary, the rook has possibilities of checking on the fifth rank as well as on the three files behind the white pieces. 85 ♖c5 85 h6 is best answered by the strong waiting move 85...♖a1! (85...♖g1+? allows 86 ♔f5 and after 86...♖h1 87 ♖g7+ White wins since the king is driven to the last rank.) 86 ♖h5 (86 ♔f5 is met by 86...♖a5+ while 8a6 h7 ♖g1+ 87 ♔f5 ♖h1 is a draw as well.) 86...♔g8 87 f5 (87 h7+ ♔h8) 87...♔h7 and the position is a draw since White cannot activate the rook without losing the h-pawn. 85...♔f6 86 ♖c6+ ♔g7! Black must be wary of traps. 86...♔f7? fails to 87 ♔g5 ♖g1+ 88 ♔f5 and Black cannot prevent the king from being driven to the last rank since 88...♔g7 is met by 89 ♖g6+. 87 ♔g5 87...♖g1+! This is the main defensive idea. Black must keep the king as active as possible. 88 ♔f5 ♖a1 This is the ideal square for the rook and a common idea is to go back and forth on the a1 and b1 squares until White shows his hand. 89 ♖c7+ If 89 ♖g6+ then Black keeps the king on the seventh rank with either 89...♔f7 or 89...♔h7. 89...♔h6 90 ♖e7 ♖b1 91 ♖e8 ♔g7 92 ♖e5 ♖a1 93 ♖d5 ♖f1 93...♖b1 was more consistent with the earlier play. 94 ♖d4 ♖a1 95 ♖d6 ♖a5+ 96 ♔g4 ♖a1 96...♖b5 is also good, as stated before. 97 ♖e6 ♖g1+ 98 ♔f5 ♖a1 99 h6+ 99...♔h7! Black maintains the pressure on the h-pawn. 100 ♖d6 ♖a2 101 ♔g5 ♖g2+ 102 ♔f6 ♔xh6! The right moment to capture the pawn. 103 ♔e7+ 103...♔h7 103...♔g7 is also a draw after 104 f5 ♖e2+ 105 ♖e6 ♖f2 106 ♖g6+ (or 106 f6+ ♔g6) 107...♔h7 108 ♔f6 ♖a2. 104 f5 ♖e2+ 105 ♖e6 ♖a2 106 f6 234 Black to move According to Botvinnik, it is necessary to have an understanding of this type of position (The king can also be placed on f7 but it makes no significant difference.) in order to correctly conduct endgames with two disconnected extra pawns on one wing. 106...♖a8! The only move to hold the draw. It prevents White from setting up the desired Lucena position with his king on f8. If Black ignores this by adopting a waiting policy with 106...♖a1? White wins: 107 ♔f8 ♖a2 108 f7 ♖a8+ 109 ♖e8 (Not 109 ♔e7 ♔g7 with a draw.) 109...♖a7 110 ♖d8 (It’s also possible to move the rook along the efile (e4-e1) threatening a check on the h-file followed by ♔g8, for example 110 ♖e2 ♖a8+ 111 ♔e7 ♖a7+ 112 ♔f6 ♖a6+ 113 ♖e6 ♖a8 114 ♖e8.) 110 ...♔g6 111 ♖d6+ ♔h7 112 ♔e8 ♖a8+ 113 ♖d8. 107 ♔f7 ♔h6 An important theoretically drawn position has arisen. Black could also play 107...♖b8 with a theoretical draw since the side is long enough. 108 ♖e1 After 108 ♖e8 Black prevents the Lucena position, ♔f8 followed by f6-f7, by 108...♖a7+ (108...♖a6 is also fine.) 109 ♖e7 (109 ♔f8 ♔g6) 109...♖a8 (The simplest but 109...♖a1 110 ♔f8+ ♔g6 11 f7 ♔f6! 12 ♖e8 ♖a7 is also a draw.) 110 ♖d7 ♖b8 111 ♔e7 ♔g6 with a draw. 108...♖a7+ 109 ♖e7 ♖a8 110 ♖d7 ♔h7 111 ♖d1 ♖a7+ 112 ♔e6 ♖a6+ 113 ♖d6 ♖a8 114 ♖d4 ♔g8 115 ♖g4+ ♔f8 and a draw was agreed. If the defender understands how to manoeuvre with the king and rook there are no problems in holding the draw in accordance with the methods we have just seen. I recommend that you practice this endgame as a defender against the computer to be sure you are able to hold the draw. 235 Kling/Horwitz 1851 White to move Provided Black doesn’t move his rook away from the important sixth rank this is an easy draw. For example 1 ♔g4 ♖c6 2 ♖f4 ♖a6 3 ♖f6+ ♖xf6 4 gxf6 ♔xf6 doesn’t work for White since the remaining pawn is a rook pawn. Note that this is the winning method if both pawns are moved one square to the left. 1 ♖b8 ♖a4+ 2 ♔e3 ♖a7! Another way to draw is 2...♖a5! 3 ♖g8+ ♔h7 4 ♖g7+ ♔h8 etc. 3 ♖g8+ ♔h7 ...and there is no way for White to proceed. 236 Kasparian 1927 Black to move If the g-pawn is defended on the sixth rank it’s a win after precise manoeuvring with king and rook. Black’s best defence is 1...♖e1+. Alternatives are: a) 1...♖a6 2 ♖d7 ♖c6 3 ♔d8+ ♔f8 4 ♖f7+ ♔g8 5 ♖e7! The rook must take control of the sixth rank so as to be able to push the h-pawn to h6. 5...♔f8 6 ♔d7 ♖b6 7 ♖e6 ♖b7+ 8 ♔d6 ♖b6+ (8...♖b1 9 h6) 9 ♔e5 ♖a5+ 10 ♔f6 followed by g6-g7+ or h5-h6 with a win. b) 1...♔h6 also loses after 2 ♖d7! ♖f1 3 ♔e8! ♖e1+ 4 ♖e7 ♖c1 5 g7 ♔h7 6 ♔f8 ♖c8+ 7 ♔f7 ♖d8 8 h6 and ♖e8 next move with a win. 2 ♔d8 ♖h1! Or 2...♔g8 3 h6 ♖e6 4 h7+ ♔h8 5 ♖g5 ♔g7 6 ♖h5 ♔h8 7 g7+ or 2...♔h6 3 ♖d7 ♔xh5 4 g7 ♖g1 5 ♔e8 ♔h6 6 ♔f8 with a win. 3 ♖d7+ ♔f8 4 ♖f7+ ♔g8 5 ♖h7 ♖a1 6 ♖c7 ♖a5 7 ♔e7! ♖a6 7...♖xh5 8 ♔f6. 8 ♔d7 The key to winning this endgame is control of the sixth rank. 8...♔g7 9 ♔c8+ ♔g8 10 ♖d7 ♖c6+ 11 ♔b7 ♖e6 12 ♔c7 ♖f6 13 ♖d6 Mission accomplished! 13...♖f1 14 h6 and White wins. The world title challenger Fabiano Caruana has said that he has made a deep study of rook endings with two pawns versus one pawn. In this way he has learned a lot about rook endings in general, since positions must be assessed after every pawn exchange. The following position illuminates this very well. 237 Khaitrullin – Inarkiev Moscow 2006 White to move Common sense tells us that White should play his king to g2 to protect the g3-weakness, while preventing the black king from penetrating on the kingside. However, this way of thinking turns out to be a mistake. 46 ♔g2?? was played and it lost after 46...f4! 47 gxf4+ ♔xf4 48 ♔f2 ♖b2+ 49 ♔e1 ♔e3 50 ♔d1 ♔d3 51 ♔c1 ♔c3 52 ♖h6 ♔b3, since White cannot prevent Black from reaching the Lucena position. After 53 ♖h3+ ♔a2 54 ♖h4 ♖b1+ 55 ♔d2 b3 56 ♖a4 ♔b2 57 ♖a8 ♖a1 58 ♖b8 ♖a7 59 ♔d3 ♖d7+ 60 ♔e3 ♔c2 61 ♖c8+ ♔b1 62 ♖b8 b2 63 ♖b6 ♖a7 64 ♔d2 ♔a1 White resigned. Now let’s see how White could have saved the situation. The solution runs 46 ♔e2!! 46...f4. 46...♖xg3 47 ♖xb4 eventually leads to the famous position by Philidor which is an easy draw. 46...♔g4 is answered by 47 ♖g6+ ♔h3 48 ♖g5 when the kingside pawns will disappear one after another and this is to White’s advantage since Black’s king is out of contact with the b-pawn. 47 gxf4+ ♔xf4 48 ♔d2 ♔e4 49 ♔c2 ♖c3+ 50 ♔b2 White’s king is positioned in front of the b-pawn instead of the other way around. So there is such a big difference between playing the “logical” 46 ♔g2 and the “illogical” 46 ♔e2. 238 Djokić – Trifunović Yugoslavia 2011 White to move It’s a fairly typical mistake to promote too early and here it was easy to miss what was coming next. White played 62 f8♕? ♖xf8 63 ♖xf8 ♔b3 64 ♔d5 (64 ♖a8 a2 65 ♖xa2 65...♔xc3!! This saving move was probably missed by White. 65...♔xa2 66 ♔d5 ♔b3 67 ♔d4 leads to an elementary win. 66 ♔d5 ♔b3 67 ♖h2 c3 68 ♔d4 c2 69 ♖h1 ♔b2 and White must sacrifice the rook for the pawn.) 64...♔xc3 65 ♖f3+ ♔b2 66 ♔xc4 a2 67 ♖f2+ ♔a3! (67...♔b1? 68 ♔b3 a1♘+ 69 ♔c3 and White wins.) 68 ♖f3+ ♔b2 69 ♖f2+ ♔a3 70 ♖f1 ♔b2 and a draw was agreed. Correct was 62 ♖b5! It’s better to build a bridge with the threat of ♖f5 than play the game move which only led to a draw. 62...♖xf7 62...♔a1 63 ♖f5 ♖xf5 64 ♔xf5 a2 65 f8♕ ♔b2 66 ♕b4+ ♔c2 67 ♕a3 ♔b1 68 ♔e4 a1♕ 69 ♕xa1+ ♔xa1 70 ♔d4 and White wins. 63 ♔xf7 ♔a1 64 ♖b4 a2 65 ♖xc4 ♔b2 66 ♖a4 and the apawn drops next move leaving White with an easy win. Remember that bridge building isn’t only for the Lucena position! 239 Lasker – Rubinstein Saint Petersburg 1914 White to move The advantage of having a more active rook behind a passed pawn is that it’s sometimes possible to place the opponent in a situation where he can only worsen his position. 59 ♖f4! The rook effectively puts Black in zugzwang. 59...b4 59...d4 gives the white king free passage in the centre. Play might continue 60 ♔e4 ♖d6 61 ♖f3! White must prevent Black’s passed pawn from advancing before pushing his own one. 61...♔c4 62 f6! (This move actually works so there is no need to stop the d-pawn with 62 ♖d3 or 62 b3+.) 62...♖e6+ (The pawn race is most definitely in White’s favour after 62...d3 63 f7 d2 64 f8♕ d1♕ 65 ♕c8+ ♔b4 66 ♕c3+ ♔a4 67 ♕a3 mate.) 63 ♔f5 ♖e3 64 ♖f4! ♖e8 65 f7 ♖f8 66 ♔e6 b4 67 ♔e7 ♖h8 68 f8♕ ♖xf8 69 ♔xf8 ♔d3 70 ♔e7 b3 71 ♔d6 ♔c2 72 ♔c5 d3 73 ♔c4 d2 74 ♖f2 and White takes on d2 next move and wins with the remaining pawn. 60 b3 The black pieces must give way to the white army. 60...♖f7 60...♔c6 61 ♔d4 ♔d6 62 ♖f2 ♖h6 63 f6 ♖h4+ 64 ♔e3 ♖h3+ 65 ♔f4 ♔e6 66 f7! An important tactical point to win this position. (66 ♔g4? leads to a draw after 66...♖h8.) 66...♖h8 67 ♔g5 ♖f8 68 ♔g6 d4 69 ♔g7 with an easy win. 61 f6 ♔d6 62 ♔d4 ♔e6 63 ♖f2! 63...♔d6 63...♖xf6 64 ♖xf6+ ♔xf6 65 ♔xd5 ♔e7 66 ♔c5 ♔d7 67 ♔xb4 ♔c6 68 ♔a5 and White wins the pawn ending. 64 ♖a2 A beautiful rook manoeuvre decides the game. 64...♖c7 65 ♖a6+ ♔d7 66 ♖b6 An excellently played endgame by Lasker, showing how to use a mobile rook. First he played a waiting move with his rook and then swung it from f2 over to a2 to decide the game with the manoeuvre ♖a6 and ♖b6. 240 Keres – Botvinnik The Hague/Moscow 1948 White to move When White is a pawn down the defender must play an active game with the remaining pieces, especially the rook. Here Black’s threat is to play ...♖c3+ and pick up the a-pawn so the question is to find the most accurate way of defending it. Keres played 50 a4 which isn’t the principled continuation. According to Botvinnik the active “50 ♖a4 would not have guaranteed a certain draw after 50...♔f6 51 h4 gxh4+ 52 ♔xh4 ♔e6 53 ♔g4 ♔d6 followed by ...♔c7-b6. White would still have had to overcome a number of difficulties to achieve a draw.” Botvinnik’s suggestion can be met by 54 ♖a5!, cutting off the black king from the fifth rank, when after the further 54...♔c7 55 ♖h5! ♔b6 56 ♔f4 White has excellent drawing possibilities thanks to his active rook. Another interesting alternative, not mentioned by Botvinnik, was 50 h4!?. If White somehow manages to eliminate all the pawns on the kingside he should have a certain draw. After 50...♖c3+ (More active though is 50...gxh4+ 51 ♔xh4 ♔g6.) 51 ♔g4 gxh4 52 ♖d7+ ♔f8 53 ♔xh4 ♖xa3 54 ♔h5 ♔e8 and White’s activity compensates for the two rook pawns and will guarantee a draw. 50...♔g6 51 h4 51...♔h5 “It was essential to retain the h6-pawn, defended by the rook; then the black king would have gained freedom of movement. With the pawn on g5 the king is tied down, and White can easily gain a draw.” Botvinnik suggests 51...gxh4 but after 52 ♔xh4 Black has two rook pawns which shouldn’t be enough for a win if White plays actively with his king and rook. 52 hxg5 hxg5 241 White to move 53 ♖d3? “White considers himself obliged to defend the third rank and he loses the game. Meanwhile 53 ♖d5! would have led to a draw, since both black pieces are tied to the defence of their pawns.” This is indeed so. White plays too passively with the rook and that is normally suicide in rook endings. 53...♖c4! With a pawn to the good and his more active rook, Black is winning. 54 ♖a3 a5 55 ♔h3 ♖b4 56 ♔g3 ♖f4 57 ♖a1 ♖g4+ 58 ♔h3 ♖e4 59 ♖a3 ♔g6 60 ♔g3 ♔f5 61 ♔f3 ♔e5 62 ♔g3 ♖d4 63 ♖a1 ♔d5 64 ♖b1 ♖b4 65 ♖f1 ♔e4 66 ♖e1+ ♔d4 67 ♔h2 Botvinnik writes that the last hope was 67 ♔f3 upon which he had planned 67...g4+ 68 ♔xg4 ♖xa4 followed by 68...♖c4 leading to a win. The problem is that White has a draw with 69 ♔f4, for example 69...♖c4 70 ♖d1+ ♔c5 71 ♔e3. The king is too close to the a-pawn and so it’s a certain draw. Instead of 67...g4+?? he should keep the pawn and play 67...♔c3 68 ♖g1 g4+ with an easy win. 67...♖xa4 68 ♖g1 ♖c4 69 ♖xg5 a4 70 ♔g2 ♔c3 71 ♔f3 a3 72 ♖a5 ♔b3 White resigned. 242 Alekhine – Capablanca World Championship match, Buenos Aires 1927 White to move This famous position from the 34th World Championship game between Alekhine and Capablanca is a text book demonstration of how to exploit an extra outside passed pawn in a rook ending where the stronger side’s rook is behind the passed pawn. Alekhine played 54 ♖a4! This move is the first step in a six-stage-plan to win this position. By positioning the rook behind the passed pawn Black is forced to place his rook in front of the pawn thereby reducing it it too passivity. 54 ♖d5 is possible and in fact Alekhine played it on move 52 although it’s not the strongest location for the rook. Tarrasch once said that the rooks belong behind passed pawns, whether they be attackers or defenders, and his statement is certainly applicable to this position. 54...♔f6 55 ♔f3 The second step is to move the king towards the centre. 55...♔e5 56 ♔e3 56...h5 Black is in zugzwang and forced to move a pawn on the kingside, despite the fact that it will be easier for White to attack on that flank in the future. Black cannot move the rook since 58...♖a7 only improves the passer after 58 a6, while 57...♔f5 allows 59 ♔d4 and a subsequent attack on the rook. Lastly 58...♔d5 gives White the opportunity to start an attack on the kingside with 59 ♔f4. 57 ♔d3 Step three involves an attack on the rook with the king. If White doesn’t obtain that possibility he goes for step four which is an attack on the kingside pawns. 57...♔d5 58 ♔c3 ♔c5 59 ♖a2! Black is in zugzwang for the second time in this endgame. 59...♔b5 Now we vary from the actual game and give the best play for both sides. 60 ♔d4! The introduction to step four. 60...♖d6+! The best defence is for the rook and king to exchange places on the blockading square a6. 61 ♔e5 ♖e6+ 62 ♔f4 ♔a6 63 ♔g5 The king is on its way to g7 to put pressure on the weakest link in Black’s pawn chain, the f7-pawn. 63...♖e5+ 64 ♔h6 ♖f5 65 f4 Before playing the king to g7 it’s useful to stop unnecessary pawn exchanges on the kingside by first playing h2-h4 to prevent ...g6-g5. 65...♖d5! 66 ♔g7 ♖d7 67 ♔f6 f7 is well defended so White concentrates on the g6-pawn instead. 67...♖c7 68 f5! White plays with a meticulous technique and breaks up Black’s pawn chain. Another more prosaic win was 68 ♖e2 ♔xa5 69 ♖e7 ♖c2 70 ♔xf7 ♖xh2 71 ♔xg6 ♖g2 72 ♖e5+! ♔b6 73 ♖g5 and White picks up the h5-pawn the next move. 68...gxf5 70...♖c6+ 71 ♔xf7 gxf5 72 ♖f2 leads to a position where all Black’s pawns will fall like ripe fruit. 69 ♔xf5 Black now has two serious pawn weaknesses. 69...♖c5+ 70 ♔f6 ♖c7 71 ♖f2! Step five in the six-stage-plan is to sacrifice the a-pawn in exchange for both the remaining black pawns. 71...♔xa5 72 ♖f5+ ♔b6 73 ♖xh5 ♔c6 74 ♖h7 ♔d6 75 ♖xf7 ...and Black can resign – without waiting for the humiliating step six: promotion of the pawns. 243 Kantorovich/Steckner/Poghosyan 2011 Black to move Until 2011 this endgame was thought to be winning, according to an analysis by Steckner. But Kantorovich and Poghosyan always believed that it was a draw – and they are right! This is how to achieve it. 1...g5! Steckner’s suggestion 1...♖xf2? loses to 2 ♖c7! ♖a2 3 a7 ♔f5 4 ♔c4!!. The idea is that after 4...♔g4 5 ♔b3 White is not only threatening the rook but also ♖c4+ followed by ♖a4 with a win. 2 hxg5+ 2 ♔d5 g4!! is a draw. 2...♔xg5 3 ♔c5 ♔g4 4 ♔b5 ♖xf2 5 ♖a8 ♖b2+ 6 ♔c4 ♖a2 7 ♖g8+ ♔f3 8 ♔b5 ♖b2+ 9 ♔c6 ♖a2 10 ♔b7 ♖b2+ 11 ♔a8 f5! 12 ♖g5 f4! 13 gxf4 h4! 14 f5 ♔f4 15 ♖h5 ♔g4 16 ♖h8 ♔xf5 17 ♖xh4 ♔e6 The position is a well-known draw. The key idea to remember is to be very active on the kingside, especially with the g-pawn. It’s possible to study this complicated rook ending in more depth by having a look in Dvoretsky’s Endgame Manual. 244 Lasker – Levenfish Moscow 1925 White to move This position looks on the surface to be a loss for White but Lasker actually missed a miraculous draw. Here he could have carried out an amazing breakthrough 49 f5!!. In the game Lasker actually played 49 ♔f6? and lost a vital tempo as well as the game. 49...exf5 49...gxf5? loses after 50 h5, for example 50...f6+ (Activating the rook with 50...♖b7 is to no avail after 51 h6 ♖b1 52 h7 ♖g1+ 53 ♔xf7 ♖h1 54 ♔g7 ♖g1+ 55 ♔f6 ♖h1 56 ♔g6 ♔b5 [56...♖g1+ 57 ♔h5 ♖h1+ 58 ♖h4 and White wins by building a bridge.] 57 ♖d4 a4 58 ♖d8) 51 ♔xf6 f4 52 h6 f3 53 ♔g6 f2 54 ♖f4 ♔d5 55 h7 ♔xe5 56 ♖xf2 (Not 56 h8♕+? ♔xf4 with a draw.) 56...♖a8 57 ♖a2 ♔d4 58 ♖xa5 ♖h8 59 ♔g7 ♖e8 60 ♔f6 ♖h8 (60...e5 61 ♖xe5 ♖xe5 62 h8♕ leads to the notorious endgame queen versus rook.) 61 ♖a4+! ♔d5 62 ♔g7 ♖e8 63 h8♕ ♖xh8 64 ♔xh8 e5 65 ♔g7 The black king is cut off so White wins easily with rook against pawn after 65...e4 66 ♔f6 e3 67 ♔f5 e2 68 ♖e4. 50 e6! This other pawn breakthrough secures White a distant passed pawn supported by the king and it’s such a strong asset that it leads to a forced draw. 50...fxe6+ 51 ♔xg6 ♔b5 52 ♖a1 f4 53 h5 e5 53...f3 doesn’t change the verdict after 54 ♖f1! a4 55 ♖xf3 a3 56 ♖f1 a2 57 ♖a1 ♔c4 58 h6 ♔b3 59 h7 ♖a8 60 ♖h1! ♔b2 61 ♖h2+ ♔b3 62 ♖h1. 54 ♖e1! The only move to draw so White must be careful. 54...♔c4 54...a4 55 ♖xe5+ ♔c6 56 ♖e4 a3 57 ♖xf4 a2 58 ♖f1 a1♕ 59 ♖xa1 ♖xa1 60 h6. 55 ♖xe5 ♔d3 56 h6 f3 57 h7 ♖xh7 58 ♔xh7 f2 59 ♖f5 ♔e3 60 ♖f8 a4 61 ♖e8+ ♔f3 62 ♖f8+ ♔g2 63 ♖g8+ ♔h3 64 ♖f8 Now it’s a clear draw. It’s a pity Lasker missed this beautiful possibility but it only goes to show that rook endings are tricky even for the best players in the world. 245 Hansen – Seitaj Olympiad, Thessaloniki 1984 White to move The winning plan is to attack the h6-pawn by the manoeuvre ♔f3-g4-h5 followed by f4-f5-f6. Black cannot do anything about this since ...f6 weakens the g6-square. 34 ♔f3 ♔f8 Black has to take care of two weaknesses. First he must control the passed pawn and secondly he must prevent White’s potential attack on the h6-pawn. This is of course an impossible mission. 35 ♔g4 ♔e7 35...g6 weakens the f6-square and White can exploit it after 36 c7 ♖c8 37 ♔f4! ♔e7 38 ♔e5 and Black is in zugzwang. 36 c7 ♖c8 37 ♔h5 37...♔f6 Transition to a pawn ending after 37...♔d7 38 f4 ♖xc7 39 ♖xc7+ ♔xc7 leads to a loss after 40 f5 ♔d6 41 f6 gxf6 42 ♔xh6 ♔e5 43 ♔g7 ♔f5 44 ♔xf7. 38 ♖c6+ ♔e7 38...♔e5 is answered by 39 g3 and f4. 39 f4 ♔d7 40 ♖c2 ♔e6 41 g4 ♔f6 42 ♖c6+ ♔e7 43 f5 ♔d7 44 ♖c2 ♔e7 45 h3 Black is in zugzwang. 45...♔f8 46 f6 ♔g8 47 ♖c6 ♔h7 48 fxg7 ♔xg7 49 g5! White wins by getting a new passer on the other side of the board. 49...hxg5 50 ♔xg5 f6+ 51 ♔f5 ♔f7 52 h4 Black resigned. A beautifully played rook ending by White. Note how helpless Black was when White tried to provoke weaknesses on the kingside. This ending is reminiscent of the next example which follows the same principles, although here Black’s rook is more passive. 246 Rubinstein – Lasker Saint Petersburg 1909 White to move When you have an extra pawn and more active pieces you can follow Rubinstein’s multi-step-plan. The first thing to do is to immobilise the black king as much as possible. 28 ♖a6! ♔f8 Black cannot liberate his rook by walking to b8 since White will decide the game with his king and passed pawn. The next step in Rubinstein’s plan is to activate the king and pawn as much as possible. 29 e4 ♖c7 30 h4! White seizes space on the kingside before advancing with the passed pawn. This also represents a defence against a future ...♖c2, as it eliminates the danger of having unprotected pawns on the second rank when White moves up with his king. 30...♔f7 31 g4 ♔f8 32 ♔f4 ♔e7 Black can only wait and avoid weaknesses. 33 h5 It is important to realise that White cannot win with just his epawn. It’s of vital importance to provoke another weakness on the kingside, in accordance with the principle of two weaknesses. 33...h6 Black couldn’t avoid a weakness on the kingside in the long run. Now the last stage in Rubinstein’s plan is the exploitation of the weakness just created on g6. If Black had continued his waiting strategy White would nevertheless have forced a weakness after, say, 33...♖b7 34 g5 ♖c7 35 e5 ♖b7 36 ♔f5 ♖c7 37 g6. This advance secures the control of the important f7-square. 37...h6 38 a4 ♖b7 39 ♖e6+! ♔d7 (If 39...♔f8 40 ♖c6 ♔e7 41 ♖c8 and Black cannot prevent ♖g8 next move. This is one reason why control of the f7square is so important.) 40 ♖f6!! Such tactical finesses are always a reality with far advanced pawns supported by the king. 41...♔e8 (41...gxf6 leads to a speedy loss after 42 g7 ♖b8 43 exf6 and f7 next move.) 42 ♖f7! Once more we come to realise the huge importance of the f7-square, deep in enemy territory. 42...♖xf7+ 43 gxf7+ ♔xf7 44 e6+ ♔e7 45 ♔e5! and White wins. 34 ♔f5 ♔f7 35 e5 ♖b7 36 ♖d6 36...♔f8 36...♖c7 is answered by the pseudo-sacrifice 37 ♖d7+! ♖xd7 38 e6+ ♔e7 39 exd7 ♔xd7 40 ♔g6 and White wins the pawns on the kingside. 37 ♖c6! 37 ♖d8+ ♔e7. 37...♔f7 38 a3! White avoids a possible ...♖b4 and places Black in zugzwang. Lasker resigned the game. A possible continuation was 38...♔f8. If 38...♖e7 39 e6+ ♔g8 40 ♔g6 ♖e8 then 41 e7! followed by ♖d6 and ♖d8. Or 38...♔e7 39 ♔g6 ♔f8 40 ♖c8+ ♔e7 41 ♔xg7 and Black’s position falls apart. 39 ♔g6 ♖b3 40 ♖c8+ ♔e7 41 ♔xg7 ♖xa3 42 ♔xh6 It’s understandable that Lasker didn’t want to endure this massacre. 247 Duras – Capablanca New York 1913 Black to move Capablanca played 38...h5 but more precise was 38...g5! which prevents White from exchanging his h-pawn. Black then increases his chances of advancing down the board. Note that the black rook is excellently placed on White’s fourth rank. If not, White to move would have played 38 h4! followed by g2-g3 setting up the most resistant pawn formation when playing with three pawns against four. The pawn wave f2-g3-h4 is a very effective defensive formation in other endings as well, apart from knight endings where the formation h2-g3-f4 is more advantageous. 39 g3! It’s necessary to stop Black from playing his pawn to h4, fixing the g2-pawn. 39...h4 It’s too early for 39...g5 due to 40 ♖a5 f6 41 ♖a7! and the king is cut off on the last rank. 40 gxh4? Correct was 40 g4 with the aim of exchanging the g-pawn for the stronger f-pawn. The weak square on f4 isn’t so serious and White will pressurise the weakest pawn in Black’s position to make it more difficult for Black to activate his king. 40...♖xh4 After this advantageous exchange it will be easier for Black to create a passed pawn on the e-file. 41 ♔g2 e5 41...♖d4! 42 ♔g3 g5! was the strongest continuation, not allowing the h-pawn to advance any further. After 43 ♖a5 ♖d5 Black plans the advance of the king to g5 followed by ...f5. 42 ♔g3 ♖d4 42...g5 plays into White’s hands after 43 ♖a5 f6 44 ♖a7. 43 ♖a5? Rabinovich recommends the continuation 43 h4! f6 44 ♖b7 ♔g8 45 ♖a7 ♔h7 46 ♖b7 ♔g6 47 ♖e7 ♖d3+ 48 ♔g2 ♖d5 49 ♔g3 f5 50 ♖e8 ♖d3+ 51 ♔g2 e4 52 ♖h8! f4 53 h5+ ♔g5 54 ♖e8 ♖d4 55 ♖e7 ♔f6 56 ♖e8 ♔f5 57 ♖f8+ with drawing chances. This is indeed so because after the further 57...♔g4 58 ♖f7 ♖d2 59 ♖xg7+ ♔xh5 60 ♔f1 it’s an elementary draw. 43...f6 44 ♖a7 ♔g8 45 ♖b7 ♔h7 46 ♖a7 ♔g6 47 ♖e7 ♖d3+ 48 ♔g2 ♖d5 49 ♔g3 f5 50 ♖a7 ♖d3+ 51 ♔g2 e4 52 ♖a4? 52 h4! was the principled move. If the h-pawn can be pushed it should be pushed since White benefits from an exchange. 52...♔g5! Black has conquered a lot of space but the position is still a draw with precise play. 53 ♖a5 g6 54 ♖b5 54 ♖a6 loses after 54 ♖d2! 55 ♖e6 (55 ♖a3 f4, 55 ♔f1 ♖d1+ or 55 ♔g2 f4) 55...♔f4. 54...♔f4 55 ♖a5 55 ♖b6 ♖d2. 55...♖d2 56 ♖a4 g5 57 ♖b4 ♔e5 58 ♖b5+ ♖d5 59 ♖b8 59...f4? 59...♖d1 60 ♖e8+ ♔f6 61 ♖f8+ ♔e6 62 ♖e8+ ♔f7 63 ♖e5 ♔f6 64 ♖e8 ♖e1! and there is nothing White can do against ...f4 according to Smyslov/Levenfish. This plan was adopted in the rook ending between Capablanca and Yates in Hastings 1930 so it seems that Capablanca learned it from the actual ending. 60 ♖g8? 60 ♔f1! ♖d1+ 61 ♔e2 was correct with drawing chances. It’s true that the pawn on h3 is weak but so are also the pawns on e4 and g5. 60...♔d4? Smyslov/Levenfish suggest the continuation 60...♖d2 61 ♖xg5+ (61 ♔f1 f3 62 ♔e1 ♖e2+ 63 ♔f1 ♖a2 64 ♔e1 ♖a1+ 65 ♔d2 ♖f1) 61...♔f6 62 ♖g4 ♔f5 63 ♔f1 f3 64 ♔e1 ♖e2+ 65 ♔f1 ♖a2 66 ♔e1 ♖a1+ 67 ♔d2 ♖f1 68 ♔e3 ♖e1+ 69 ♔d4 ♖e2 70 ♖g8 ♖d2+ 71 ♔e3 ♖d3 mate. 61 ♔f1 ♔d3 62 ♖a8 e3 Smyslov/Levenfish analyse three alternatives: a) 62...f3 63 ♔e1. b) 62...♔d2 63 ♖e8 ♖d4 64 ♖e5 ♖b4 65 ♖d5+ ♔c3 66 ♖xg5 ♖b1+ 67 ♔e2 f3+ 68 ♔e3 ♖e1+ 69 ♔f4. c) 62...♖b5 63 ♖d8+ ♔c4 64 ♖c8+. 63 ♖a3+? White is now lost. The last chance to hold the draw according to Smyslov/Levenfish was 63 ♖e8! ♔d2 64 ♖e7 ♖d3 65 ♖e5 ♔d1 66 ♖e8 exf2 67 ♖e5 ♔d2 68 ♔xf2 but not 68 ♖xg5?? ♔e3. 63...♔e4 64 fxe3 64...f3! 64...fxe3? 65 ♖a8 ♖f5+ 66 ♔e2 ♖f2+ 67 ♔e1 ♔f3 68 ♖f8+ ♔g3 69 ♖g8 (69 ♖e8 is suggested by Smyslov/Levenfish but it’s actually a mistake after 32...♔xh3! 33.♖xe3+ ♔g2 and Black wins.) 69...♖f5 70 ♖h8. 65 ♔g1 Or 65 ♔f2 ♖d2+ 66 ♔g3 ♖g2 mate. 65...♖d3 66 ♖a8 66 ♖a5 ♔xe3. 66...♔xe3 67 ♖e8+ ♔f4 68 ♖g8 ♖d1+ 69 ♔f2 ♖d2+ 70 ♔f1 70 ♔g1 g4! is a typical idea to remember! 71 ♖xg4+ (71 hxg4 ♔g3) 71...♔e3 Smyslov/Levenfish 72 ♖g8 (72 ♖g3 ♔e2) 72...♖d1+ 73 ♔h2 f2 74 ♖e8+ ♔d2 75 ♖f8 f1♕ 76 ♖xf1 ♖xf1 with an easy win. 70...♖h2 71 ♔g1 ♖xh3 72 ♖g7 g4 73 ♖g8 ♔g3 White resigned. 248 Duras – Nimzowitsch San Sebastian 1912 Black to move The reason we are discussing this complicated endgame again is that when we looked at Duras – Capablanca we learned how to win and defend technically. However, there is another way to win if technique alone isn’t enough and that is with the help of psychology. This is what Nimzowitsch managed to do against the same opponent of Capablanca, and in the same type of ending! Compared with the game against Capablanca, Duras had an improved position – but he still lost. How that happened can be useful to know from the perspective of both the winner and the loser. Duras has managed to set up the strongest defensive pawn wave formation h4-g3-f2 which is the safest way to head for a draw when playing with three pawns versus four on the same wing. How can Black play for a win in such a situation? Nimzowitsch tried to make his opponent tired and careless by doing nothing. As a matter of fact there is nothing Black can do. By playing in this “innocent” manner Nimzowitsch managed to fool his opponent. Let’s see how he did it. Nimzowitsch played 33...♖a5 ...which was met by 34 ♔g2. However, the strongest move was 34 f4! preventing Black from activating his e-pawn. In this position the absolute seventh rank means nothing since Black is unable to activate his king without losing a bunch of pawns on the kingsde. A plausible continuation would then be 34...♖a2 35 ♖c7 g6 36 ♖b7 ♔f5 37 ♖xf7+ ♔g4 38 ♖f6 ♔xg3 39 ♖xg6+ ♔xf4 40 ♖xe6 ♔g3 41 ♖e3+ with a draw. 34...♖a8 Nimzowitsch plays a patient and demanding waiting game. Technically 34...e5 is the logical move, preparing the activation of Black’s pawn position as much as possible and then trying to squeeze White to death. Now a lot of nothing-is-happening moves are played until move 56: 35 ♔f3 ♖e8 36 ♖b5 ♔g6 37 ♖g5+ ♔h6 38 ♖a5 ♖e7 39 ♖b5 ♖a7 40 ♖c5 ♖a3+ 41 ♔e2 ♖a7 42 ♔f3 ♖b7 43 ♖a5 ♖b3+ 44 ♔e2 ♖b7 45 ♔f3 ♖c7 46 ♖b5 ♖c3+ 47 ♔e2 ♖c7 48 ♔f3 ♖d7 49 ♔e3 ♖d1 50 ♖a5 ♔g6 51 ♖g5+ ♔h6 52 ♖a5 ♖b1 53 ♖c5 ♖a1 54 ♖b5 ♖a4 55 ♖c5 ♖a8 56 ♖b5 56...♖h8!! Very clever play by Nimzowitsch who had strong fighting spirit as well as high-level psychological capabilities. On a superficial level Black’s plan is apparently to manoeuvre his king to f6 after which the h5-pawn must be defended. Duras’ priority in this situation seems to have been the prevention of the king reaching f6. 57 ♖g5?? Duras wanted to limit the activity of Black’s king but forgot what Black could do with White’s rook! 57...f5! Wow! One can just imagine Nimzowitsch’s euphoria when he trapped the white rook. Black’s winning plan is to manoeuvre the rook to g4 and force a pawn ending where the extra pawn can be exploited for a win. 58 ♔f4 58 f3 ♖a8 59 g4 hxg4 60 fxg4 g6 61 gxf5 certainly liberates the white rook but after 61...gxf5 Black still wins thanks to his connected passed pawns. 58...♖a8 59 ♔e5 ♖a6 White is in zugzwang since the pressure on e6 cannot be maintained. 60 ♔f4 ♖a4+ 61 ♔e5 61 ♔f3 ♖g4 62 ♖xg4 hxg4+ 63 ♔f4 ♔g6 64 ♔e5 ♔f7 and Black wins the pawn ending. 61...♖e4+ 62 ♔d6 g6 63 f3 63 ♔e7 is obviously answered by 63...♔g7!. 63...♖e3 64 g4 fxg4 65 fxg4 ♖g3 66 gxh5 ♖xg5 67 hxg5+ ♔xh5 White resigned. This endgame has never been kind to Duras who lost twice in a painful way instead of achieving two solid draws. In particular, the endgame against Nimzowitsch must have been a nightmare but all the same it showed what a skilful psychological player Nimzowitsch was. He simply hypnotised his opponent! Such ability is an important component in habitually winning games, or turning half points into full points. If Duras can be manipulated like this, then the moral of the story is that it could easily happen to other chessplayers too. Remember that this psychological method of inducing your opponent to lose concentration can be used in other “dry” positions. But the secret is to find an idea which involves something similar to the imaginative 56...♖h8(!!) and that includes all other subtle moves before the crucial move itself is actually played as well as perfect timing when to snare the opponent. 249 Eliskases – Bogoljubow Match 1939 White to move Here it is actually preferable to have a doubled pawn on f3 rather than on the normal g2 or g3 squares. This is because there are no profitable pawn exchanges for Black. Another advantage of the damaged pawn structure is that the king can be placed actively on g3. However, one drawback is the open g-file which might be exploited by Black. 27 ♖a1 g5 Znosko-Borovsky suggests the interesting idea of placing the rook on g5 and forcing the king to declare its intentions. If White places it on f1 Black can strive for the following set-up: pawns on f6 and e5 and then to advance the king to h3, targeting the weak h2pawn. If White tries to stop the king when it is placed on h5 by placing the rook on the fourth rank, Black can advance the f-pawn to f4 blocking the rook. In that position the black rook defends both the g7- and e5-pawns and Black’s king can advance to h3 and pick up the h2-pawn. This is probably the best practical winning chance for Black. If White meets the check on g5 with ♔h1 Black can try to reach the other side with his king and focus his attention on the f-pawns or a rook exchange. A plausible variation is 27...♖b5 28 ♔g2 ♖g5+ 29 ♔h3. This is one way to counteract Black’s plan and more active than 29 ♔f1 or 29 ♔h1 and will increase his drawing chances. The game continued 28 h3 ♔g7 29 ♖a5 ♔g6 30 ♔g2 h5! Black wants to get rid of the h-pawn sometime in the future and prepares a timely ...g5-g4, even though it’s also in White’s interest to increase his drawing chances with pawn exchanges. 31 ♖c5 f6 32 ♖a5 ♖b3 33 ♖c5 e5 34 ♖c6 ♔f5 35 ♖a6 ♖b4 Of course not the suicidal 35...♔f4?? 36 ♖xf6 mate. 36 ♖c6 ♖f4 37 ♖a6 g4 Black cannot improve his position any further and changes the pawn structure. 38 hxg4+ hxg4 39 fxg4+ ♖xg4+ 40 ♔f3 The position has crystallised to a pretty clear draw. 40...♖b4 41 ♖a3 ♔g5 42 ♖e3 ♖d4 43 ♔g2 e4 44 ♖e2 ♔f4 45 ♖a2 f5 46 ♖a8 ♖d2 47 ♖e8 ♖e2 48 ♖a8 ♖d2 48...e3 doesn’t lead to anything substantial after 49 ♖a4+ ♔g5 50 ♔f3 ♖xf2+ 51 ♔xe3 with a theoretical draw. 49 ♖e8 ♖d7 50 ♖a8 ♖d5 51 ♖e8 ♖d2 52 ♖e7 ♖d8 53 ♖a7 ♔g5 54 ♖a5 ♔f6 55 ♔f1 f4 56 ♔e2 ♖b8 57 ♖a6+ ♔f5 58 ♖a5+ ♔g4 59 f3+! A good finesse to remember. It’s an easy draw when Black has two doubled pawns on the f-file. 59...exf3+ 60 ♔f2 ♖b2+ 61 ♔f1 ♖h2 61...♔g3 62 ♖g5+. 62 ♖b5 ♖h5 63 ♖xh5 ♔xh5 64 ♔f2 The pawn ending is also a draw when the pawns are as far advanced as here. 64...♔g4 65 ♔f1 ♔f5 66 ♔f2 ♔e4 67 ♔f1 ♔e3 68 ♔e1 A draw was agreed. Compared to Znosko-Borovsky’s ambitious plan, Bogoljubow’s play was rather uninspired. This is an excellent position to play out with the computer – with both colours! The ending is a theoretical draw but since it involves some pitfalls if the attacker plays accurately, it’s worth investigating more deeply. It’s a rather common rook ending in practical play. I had the doubled pawns myself once and managed to draw with a significantly higher rated player, but of course I already knew the above mentioned endgame from one of Euwe’s endgame books! 250 Spielmann – Rubinstein Saint Petersburg 1909 Black to move This famous example, featuring Rubinstein in his prime playing the leading role, is illuminating since it shows how to play against weak and isolated pawns with the help of optimal activity and cooperation by the rook and the king. In this position White has four pawn islands and Rubinstein puts early pressure on two of the pawns, far away from the defending king. 40...♖a8! 40...♖b3 is the wrong way to play since White is far too active after 41 ♖a2 ♖d3 42 a4! ♖xd4 43 a5 ♖c4 44 a6 ♖c8 45 a7 ♖a8 46 ♔g3 and has at least a draw. 41 ♖c3 If 41 ♖a2 ♖a4 42 ♔g3 then 42... ♔e7! since activity is the most important principle in rook endings. After 43 ♔f3 ♔e6 44 ♔e4 g5 Black rules. 41...♖a4! 42 ♖d3 ♔e7 The position of the black rook cannot be improved so it’s time to activate the king. 43 ♔g3 43 d5 is met strongly by 43...g5! followed by ...♖a7 and ...♔f6 with activity. Note that the immediate 43...♔f6? is answered by 44 ♖f3+. 43...♔e6 44 ♔f3 ♔d5 45 ♔e2 A better defence was 45 h4! with the deadly trap 45...♖xd4?? (45...♔c4 would maintain the positional pressure.) 46 ♔e3 ♖xd3+ 47 ♔xd3 and White wins the pawn ending thanks to the outside passed pawn. 45...g5! Rubinstein fixes the h3-pawn which becomes a future target. 46 ♖b3 46...f6! Rubinstein adopts a classical method whereby three tactical weaknesses on the kingside become one. Not 46...♖xd4? 47 ♖b5+ and the g5-pawn is lost next move, nor 46...♔xd4 47 ♖g3! with the double threat ♖g4+ and ♖xg5. 47 ♔e3 47 ♖b7 ♖xa3 48 ♖xg7 ♔xd4 or 48...♖xh3 with good winning chances for Black. 47...♔c4 48 ♖d3 d5! The d4-pawn is fixed. 49 ♔d2 ♖a8! 50 ♔c2 50 ♔e3 ♖e8+ 51 ♔d2 ♖e4 and White cannot prevent one of his pawns being lost. 50...♖a7! Black waits for the king to move to d2. 51 ♔d2 51...♖e7! Apparently there is no threat but more seriously White is now in zugzwang. 52 ♖c3+ 52 ♔c2 ♖e2+ 53 ♖d2 ♖xd2+ 54 ♔xd2 ♔b3! and Black wins thanks to his active king. 52 ♖e3 ♖xe3 53 fxe3 f5 54 ♔c2 ♔b5 and White cannot simultaneously control the a-pawn and Black’s passed f-pawn. 52...♔xd4 53 a4! ♖a7 54 ♖a3 ♖a5! Black blockades the pawn as quickly as possible. 55 ♖a1 ♔c4 56 ♔e3 d4+ 57 ♔d2 ♖f5! 58 ♔e1 58 a5 ♖xf2+ 59 ♔e1 ♖b2 60 a6 ♖b8 61 a7 ♖a8 followed by the deciding manoeuvre to b6 with the king winning the pawn. 58...♔b4 59 ♔e2 ♔a5 60 ♖a3 ♖f4! A very good square for the rook. It defends the d4-pawn while putting pressure on the f2-pawn. 61 ♖a2 61...♖h4! This is another benefit of the f4-square, which works as a pivot for further manoeuvring against enemy weaknesses. 62 ♔d3 62 ♖a3 ♔b4 and the white rook is overloaded by having to protect both rook pawns. 62...♖xh3+ 63 ♔xd4 ♖h4+! 64 ♔d3 ♖xa4 65 ♖e2 ♖f4! 66 ♔e3 ♔b6 67 ♖c2! ♔b7 Black must prevent ♖c8 followed by ♖g8. 68 ♖c1 ♖a4 Black’s rook is on its way to c8. 69 ♖h1 ♔c6 70 ♖h7 ♖a7 71 ♔e4 ♔d6 72 ♔f5 72...g6+! A nice transition to a winning pawn ending. 73 ♔xg6 ♖xh7 74 ♔xh7 ♔e5 75 ♔g6 g4! White resigned. Since it’s impossible to prevent Black from queening in the near future. 251 Capablanca – Alekhine New York 1924 Black to move White has an extra pawn but but he has three isolated pawns versus Black’s two. Black has the more active king and rook so there are good drawing chances for him despite his material deficit. Alekhine played 37...♔e6. The simplest way to draw was to activate the rook to its optimum square. The continuation 37...♖b4! 38 ♔d3 c5 39 ♔c3 ♖a4! features a strong rook manoeuvre which prevents White from playing the desirable a2-a3. It’s hard for White to improve his position any further since Black’s rook is just too strong. This isn’t the first time we see the importance of an active rook. 38 ♔e3 c6 38...♖b4 is no longer so strong since it allows White’s king too much activity after 39 ♔d4 ♖a4 40 ♔c5. Another possibility 38...c5 39 ♖d5 ♖b2 was probably not Alekhine’s cup of tea but it looks playable considering that it leads to further pawn exchanges. 39 h4 39 c5, exploiting the d6-square, might have been answered by the active 39...♖b5 40 ♖d6+ ♔e5 41 ♖xc6 ♖a5 and Black has fine compensation for the two minus pawns. The main threat is ...♖a3+. Another move is 39 h3 with the idea 39...♖h8? 40 c5!. 39...♖h8! Black gains a tempo for a manoeuvre to the fifth rank. 40 g3 ♖h5 Black’s rook is perfectly placed on the fifth rank, both for defence and attack. 41 ♖h2 ♖a5 42 ♔f4 f6! 43 ♖c2 White plans c4-c5, cutting off the black rook, followed by g3-g4 and h4-h5. Not immediately 43 g4? since Black has the effective reply 43...g5+!. 43...♖e5 44 c5 ♖h5 45 ♖c3 White plans ♖a3 in the hope of exchanging the c-pawn for the apawn but it’s easily prevented by Alekhine’s next move. 45...a5! 46 ♖c2 ♖e5 47 ♖c3 ♖h5 48 ♔f3 ♔e7 49 ♔g4 ♔f7! 50 ♖c4 50 ♔h3 g5 51 ♔g4 ♔g6, followed by 52...gxh4 and 53...♖e5 with pressure on White’s pawns, shows the deeper point of 51...♔f7!. 50...♔g7 51 ♖d4 ♖xc5 52 ♖d7+ ♔f8 53 ♔f4 ♔g8 54 ♖a7 ♔f8 55 a4 ♔g8 56 g4 56 ♔e3 ♖c3+ 57 ♔d4 ♖xg3 58 ♖xa5 ♔f7 59 ♖a8 ♖a3 is another draw. 56...g5+! 57 hxg5 ♖xg5 58 ♖a6 ♖c5 59 ♔e3 ♔f7 60 ♔d4 ♖g5 61 ♖xc6 ♖xg4 62 ♖c5 ♖g5 Here the players agreed to a draw. Alekhine gives 63 ♖xg5 fxg5 64 ♔e5 ♔g6 65 ♔d6 ♔f7 with a draw. The isolated pawns were a dead weight in White’s position. They had no offensive power and confined the white rook to passivity . 252 Flohr – Vidmar Nottingham 1936 White to move White has a small positional advantage consisting of his better pawn structure (two pawn islands versus three) and stronger rook and king. Despite these significant advantages it’s not so easy to convert them to a win. Firstly, you need a high level of technical precision and, secondly, it’s not possible to win against precise defence. In practice it’s not easy to defend these kinds of endings and even famous masters and grandmasters make mistakes that eventually lead to a loss. 32 ♔e2 ♔e7 33 ♔d3 ♔d6 34 ♖a5! The natural move 34 ♔d4?! helps Black to place his rook on a more active square after 34...♖b8 35 ♖a5 ♖b6 with good drawing chances. Remember that in rook endings everything is about activity. 34...♖a8 35 ♔d4 White prevents ...c5. 35 b4?! would be less effective due to 35..♔c7 36 ♔d4 ♔b6. 35...f5! A good defensive move halting White’s plans of grabbing space on the kingside with g2-g4 and h2-h4 in combination with the central push e2-e4. If White plays e2-e4 in this position Black can exchange twice. It’s a well-known defensive strategy to exchange pawns since many rook endings are drawn when material has been reduced. 36 b4 ♖b8 A more active and principled continuation was 36...♔c7 37 ♔c5 ♔b7 38 ♔d6 ♖e8 39 ♖a3 d4! 40 exd4 ♖e2. By sacrificing the pawn Black’s rook will acquire activity on White’s second rank. Black has good drawing chances after 41 ♖c3 ♖xg2 42 ♖xc6 ♖xh2 43 a4 g5!. Don’t miss any chance to activate a defensive rook by means of a pawn sacrifice! 37 a3 ♖a8 37...♖b6? would leave Black more likely to succumb to zugzwang in the near future, while also allowing White to improve his position further. 38 e4! The main idea of this central breakthrough is to penetrate the kingside. 38...fxe4 39 fxe4 dxe4 40 ♔xe4 40...♖a7? A better defence was 40...♔c7! 41 ♖e5! (41 ♔f4 ♖f8+! 42 ♔g3 ♔b6 leads to a clear draw.) 41...♔b6 42 ♖e7 a5! Black plays for maximum activity. 43 ♖xh7 axb4 44 axb4 ♖a4 45 ♖g7 ♖xb4+ 46 ♔f3 ♖h4 47 h3 ♖h6 48 ♔g4 c5 49 ♔g5 ♖h8 50 ♖xg6+ ♔b5 51 ♖g7 c4 52 h4 ♖c8 and Black has enough counterplay to achieve a draw. 41 ♔f4 h6 Waiting play doesn’t work since White improves his position with every move, for example 41...♖a8 42 ♔g5 ♖a7 43 ♔h6 ♔e6 44 g4 ♔d6 45 h4 ♔e6 46 g5 ♔d6 47 a4 ♔e6 48 b5 etc. 42 h4 ♔e6 43 ♔g4 ♖a8 43...♔f7 is answered by 44 h5 as well. 44 h5! g5 44...gxh5+ is answered by 45 ♔xh5 ♖g8 46 g4! ♔d6 47 ♖xa6. 45 g3 ♖a7 46 ♔f3 ♖a8 47 ♔e4 ♖a7 48 ♖e5+! White forces Black’s king to make a decision. If it moves to the kingside White advances his king to the queenside and vice versa. 48...♔d6 48...♔f6 also loses after 49 ♖c5 ♖c7 50 ♖a5 ♖a7 51 ♔d4 ♔e6 52 ♔c5 when it’s too late for Black to activate his rook. 52...♖d7 53 ♖xa6 ♖d3 54 ♖xc6+ ♔f7 55 a4 ♖xg3 56 ♖xh6 and Black’s position falls apart. 49 ♖e8 c5 49...♖e7+ 50 ♖xe7 ♔xe7 51 ♔e5 leads to pawn ending where White’s king is too active. 50 ♖d8+! ♔c6 50...♔c7 51 ♖h8 cxb4 52 ♖h7+ ♔b8 53 ♖xa7 ♔xa7 54 axb4 ♔b6 55 ♔f5 ♔b5 56 ♔g6 ♔xb4 57 ♔xh6 a5 58 ♔xg5 and White queens first. 51 ♖c8+ ♔b6 52 ♖xc5 ♖h7 53 ♖e5 ♔c6 54 ♖e6+ ♔b5 55 ♔f5 ♖f7+ 56 ♖f6 Black resigned. 253 Chigorin – Tarrasch Budapest 1896 White to move There is no doubt that White has the advantage with his more active pieces as well as the better pawn position but with precise play Black can still draw. It’s because of such positions we have the famous saying that all rook endings are drawn. Chigorin first improved the placement of his king. 28 ♔f3 h5 Black stops once and for all the advance of White’s h-pawn. Another idea was to cut off the king with 28...♖a4 and reply to 29 h5 by 29...h6, halting the pawn. After 30 ♖a7 play might continue 30...♖a5! 31 g4 ♖a4 (The active 31...f5! leads to a draw as well.) 32 ♔g3 ♖a2 33 f3 ♖a4 34 ♖c7 ♖b4! 35 ♖c6 ♔g7 36 ♖xa6. This is a well-known drawn position. Black is a pawn down and has a passive position but White cannot win. 29 ♖c7 ♖a5 The threat was 30 ♖c5. In this position Tarrasch offered a draw and Chigorin correctly declined since White cannot lose anyway. 30 ♔f4 ♔f8 31 f3 More precise was 31 ♖a7 with the idea of meeting 31...♔g8 with 32 f3, since then 32...♖a4+? 33 ♔f5 ♖a3 is no good because of 34 ♔g6 ♔f8 35 ♖f7+. However, it’s better to play 32...♖a3, for example, and prevent White from advancing by 33 ♔f5. 31...♔g8 A better and more active defence was 31...♖a4+ 32 ♔f5 ♖a3 33 f4 ♖xg3 34 ♔xf6 ♔g8 35 ♖c5 ♖g4 with a draw. 32 ♖a7 ♔f8 Black plays a waiting game and it actually works here. 33 g4 hxg4 Simpler was 33...♔g8. 34 fxg4 34...♖a1?? An incredible mistake by Tarrasch which decides the game. He should have continued his waiting policy with 34...♔g8!. After 35 h5 it’s possible to sacrifice the a-pawn with 35...♖b5! since 36 ♖xa6 ♔g7 37 ♖a7+ ♔g8 leads to a well-known drawn position, as showed by Friedstein. A plausible continuation is 38 h6 ♖c5 39 ♖g7+ ♔f8! (Not 39...♔h8? 40 ♖f7 ♖c6 41 ♔f5 and the f6-pawn is lost.) 40 ♖g6 (40 ♔g3 ♖c1!) 40...♔f7 41 h7 ♖c8 42 ♖h6 ♔g7 43 ♖h1 ♖a8 with a draw. But not 43...♖h8?? 44 ♔f5 ♔f7 45 ♖a1 and White wins. 35 ♔f5 ♖f1+ 36 ♔g6 ♖f4 37 g5 fxg5 37...♖xh4 loses to 38 ♖a8+ ♔e7 39 gxf6+, etc. 38 hxg5 38...♖a4 It’s impossible to set up a defensive position with 38...♔g8 due to 39 ♖a8+ (Of course not 39 ♖xa6?? ♖f8 with a theoretical draw.) 39...♖f8 40 ♖xf8+ ♔xf8 41 ♔h7. Placing the rook behind the enemy pawn doesn’t work either after 38...♖g4 39 ♖xa6 ♖g1 40 ♖a8+ ♔e7 41 ♖g8 and we have a Lucena position. 39 ♖a8+ ♔e7 40 ♔h6 a5 41 g6 ♖a1 42 g7 ♖h1+ 43 ♔g6 ♖g1+ 44 ♔h7 ♖h1+ 45 ♔g8 ♖a1 46 ♖a7+ ♔e8 47 ♖a6! ♖h1 Otherwise White decides the game with ♔h7 next move. 48 ♖xa5 The Lucena position has arisen. 48...♖e1 49 ♖h5 ♖g1 50 ♖e5+ ♔d7 51 ♔h7 Black resigned. This famous rook ending was played with precision by Chigorin. In fact it was very instructive to see how Chigorin won and how Tarrasch could have held the draw on several occasions. 254 Flohr – Petrovs Semmering-Baden 1937 White to move Superficially this position looks dangerous for the great endgame expert Salo Flohr, since Black has a rook exerting pressure on f2 and a distant passed pawn supported by his king. However, a more important asset is that White has only one pawn chain whereas Black’s is split into three parts. This endgame has the same theme as the previous Chigorin – Tarrasch game but here Black is far more active and can achieve a draw by precise play. 32 ♖a5! ♖b5?! 32...♔b6! is the strongest move, underlining that activity is more important than material. This is certainly true for this position but also for rook endings in general. Play might continue 33 ♖xd5 a5 34 ♖d8 ♔b7 35 ♖f8 a4 (35...f6 36 ♖g8 makes no difference, it’s still a draw.) 36 ♖xf7+ ♔b6 37 d5 White opens the fourth rank for the rook, which now has the possibility of attacking the a-pawn in two directions, via f8-a8 or f4-a4. After 37...a3 White plays 38 ♖f8 ♔b7 39 ♖f4 and now Black must find the only move 39...♖b6!. After 40 ♖a4 ♖a6 41 ♖b4+ ♖b6 it’s a draw. However, 41...♔c7?! is imprecise due to 42 ♖b1. 33 ♖a2! White defends the second rank. 33...g6 After 33...a5 34 ♔f4 White’s king heads for e5. The best way to prevent the white king’s ambitions on the kingside and in the centre was 33...♖b6! with the idea 34 ♔f4 ♖e6! 35 ♔g5 g6. 34 ♔f4 f6 This pawn structure is slightly weaker than the wave f7-g6-h5. 35 g4 Another idea was 35 f3 followed by g4. 35...hxg4 Of course not 35...a5?? 36 gxh5 gxh5 37 ♔f5 and White cleans up on the kingside. 36 ♔xg4 ♖b1! Black must prevent f4-f5 creating a passer on the h-file. 36...a5? 37 f4 ♔c6 38 f5 gives White good winning chances with his more active king and dangerous h-pawn. 37 ♖a5! ♖b5?! 37...♔b6! 38 ♖xd5 a5 was the best and most active option. 38 ♖a1! Now the first rank is more important than the second. 38...♖b2 39 ♔f3! Another option was 39 f4!? followed by f5 but Flohr prefers to defend the wave f2-e3-d4 and prepare ♖g1 followed by ♖xg6. 39...♖b6 40 ♖g1 f5 Black defends g6 but e5 and g5 are weakened. 41 ♔f4 ♖e6! This looks passive but it’s important to prevent the incursion White has prepared with the advance 42 ♔e5. 42 ♖b1+ 42...♔a7? Here the king is too passive. It was better to play 42...♔c7 with reasonable drawing chances. 42...♖b6? leads to a winning queen ending after 43 ♖xb6+ ♔xb6 44 ♔g5 a5 45 ♔xg6 a4 46 h5 a3 47 h6 a2 48 h7 a1♕ 49 h8♕ ♕g1+ 50 ♔f6 ♕xf2 51 ♕d8+ ♔a6 52 ♕d6+ ♔a7 53 ♕d7+ ♔a6 54 ♕xf5 and Black cannot capture on e3 due to the decisive check on e6. 43 h5? It would be stronger to avoid the check on e4 and play 43 ♔g5! followed by h4-h5. 43...♖e4+ 44 ♔g5 44... ♖g4+? Correct was 44...gxh5! 45 ♔xf5 h4 and Black’s two distant passed pawns should secure the draw. 45 ♔f6 gxh5 46 ♔xf5 White’s plan is to create two passed pawns in the central area. 46...♖g2 46...♖e4! was a better try to stop White’s plan, whereupon White must find 47 ♖e1!. 47 ♔e5 ♖g5+ After 47...♖xf2 48 ♔xd5 White’s central connected pawns are significantly stronger than Black’s disconnected rook pawns. 48 ♔e6 h4 49 ♖h1 ♖h5 50 f4 ♔b6 51 f5 ♔c7 52 f6 ♔d8 53 ♖f1! ♖h6 Or 53 ...♔e8 54 ♖b1 ♔d8 55 ♖b8+ ♔c7 56 f7. 54 ♔f7! 54 ♔xd5 also wins but it’s slower. Black resigned. A beautifully conducted rook ending by Flohr. Apart from his 43rd move it was played to perfection. It strikes one that Flohr handled the pawns according to the famous principles of Philidor, who had a great belief in pawn chains and the collective mobility of the pawns. Black had possibilities of holding the draw but that demanded more activity and above all more concrete play. This rook ending, like many others, only goes to show that rook endings are indeed difficult to play. 255 Kashdan – Alekhine Olympiad, Folkestone 1933 Black to move Black has two pawn islands and is positionally worse. It’s very instructive to see how Alekhine managed to solve his problems in the most active and dynamic fashion! 36...h5 37 ♖h7 e4 38 ♔e2 38...f5! In this situation it’s possible to sacrifice material for activity. Black parts with his least valuable pawn in order to prevent White’s king from advancing to e3. It’s possible to play in a passive manner such as 38...♖e5 39 ♔e3 ♔d5 40 h4 ♔e6 41 ♖a7 ♔d5 42 ♔f4 ♔e6 43 ♖a6+ ♔e7 and after 44 ♖a4 sacrifice with 44...e3 45 fxe3 ♔f7. However, this passive defence probably didn’t suit Alekhine who usually preferred an active defence – a golden rule in rook endings. Another way to defend was 38...♖f5 39 ♔e3 ♖f3+ 40 ♔xe4 ♖xf2 41 ♖xh5 ♔f7 but it leads to a more difficult drawn ending for Black. Whenever one has to choose between active or passive variations one should as a rule always choose the active one in rook endings! 39 ♖h6+! 39 ♖xh5?! ♖a2+ 40 ♔f1 ♖a1+ 41 ♔g2 ♖a2 42 ♖h8 ♔f7 is an easier draw for Black since ...e2-e3 cannot be effectively prevented. 39...♔e5 40 ♖xh5 ♖a2+ 41 ♔f1 41...e3! Alekhine plays for maximum activity and sacrifices another pawn. With regard to this endgame you can clearly see that rook activity is worth one pawn and king activity another pawn! This is probably valid as a general principle, at least when White’s king and rook are as passive as they are here. Note that 41...♖a1+ 42 ♔g2 ♖a2 threatening ...e3 is not effective due to 43 ♖h8!. But not 43 g4?! ♔f4!. 42 fxe3 ♔e4 43 ♔g1 If White plays 43 h4 instead, Black has an amazing continuation: 256 Black to move 43...f4!! This is a very instructive pawn sacrifice designed to prevent checks from the rear after Black advances his king in the process of playing for mate and pawn wins. 44 gxf4 (44 exf4 ♔f3 45 ♔e1 ♔xg3 46 f5 ♔g4 and Black collects another pawn with an elementary draw.) 44...♔f3 45 ♔g1 ♖a1+ 46 ♔h2 ♖a2+ 47 ♔h3 ♖a1 and White cannot escape from the continual harassment of the strongly coordinated black pieces and has to allow a draw by repetition. Amazing stuff! 43...♖e2 44 ♖h4+ ♔e5 For once Alekhine chose a passive defence but it’s also the most pragmatic move. He could have acted in accordance with the maximum-activity principle and played 44...♔xe3, not fearing that the king would be cut off after 45 ♖f4. In fact that doesn’t matter as after 45...♖e1+ 46 ♔g2 ♖e2+ 47 ♔h3 47...♖f2!! the pawn ending is a draw, e.g. 48 ♖xf2 ♔xf2 49 ♔h4 (49 g4 f4 50 g5 f3 51 g6 ♔e3 52 g7 f2 53 g8♕ f1♕+ leads to a drawn queen ending.) 49...♔g2 50 h3 ♔h2!. 45 ♖h8 ♔f6 46 ♖f8+ ♔g6 47 ♖e8 ♔f7 48 ♖c8 ♖xe3 Two pawns versus one on the same flank is normally an easy draw. The passed pawn will not change this verdict. 49 ♔f2 ♖a3 White’s king is cut off from the third rank. 50 h4 ♔f6 51 ♖c6+ ♔f7 52 ♖c2 ♖b3 53 ♖e2 ♔f6 54 ♖e3 ♖b4 Now the king is cut off on the fourth rank and it’s not easy to contend with the rook on the fourth rank. 55 ♔f3 ♖a4 56 ♖b3 ♖c4 57 ♖b6+ ♔f7 58 ♖d6 ♖a4 Drawn. White cannot win this ending since White’s king is unable to penetrate the black position. After 59 ♔e3, with the idea of ♖d4, Black replies 59...♖e4+ 60 ♔f3 (otherwise ...♖g4) 60...♖a4 etc. The white rim-pawn is worthless since after its advance Black’s king just moves in front of it and picks it up. This rook ending has some similarities to the famous one between Capablanca and Tartakower, position 258, which showed that in rook endings you must sometimes sacrifice one or even two pawns for activity. In one variation after 43 h4 f4!! we even saw a sacrifice of three pawns! So the moral is that you should never be afraid of sacrificing one pawn for activity when the great masters are ready to sacrifice two or even three in some rare cases! 257 Tarrasch – Rubinstein San Sebastian 1911 Black to move White is a pawn up and has an active rook. In addition he has a 3-1 pawn majority on the queenside. The reason Black can stay alive is White’s weakened kingside. 32...♖d8! 33 ♖a6 33 ♔e2 looks like the safer option but after 33...f4 Black’s play on the kingside leads to good counterplay. If White plays the seemingly strong 34 ♖a4 ♔f5 35 ♖d4 the pawn ending is surprisingly drawn after 35...♖xd4 36 cxd4 b5!. For example 37 a3 is met by 37...♔e6 38 ♔d2 ♔d5 39 ♔c3 h5 40 h4 ♔e6 41 b3 ♔d5 42 a4 b4+!. This is the main idea showing why this pawn ending is a draw. 43 ♔xb4?? cannot be played due to 43...♔xd4 followed by a promotion on e1 with check. 33...♖d2 34 ♖xb6+ ♔g5 35 ♔e1 ♖c2 36 ♖b5? A better winning attempt was 36 a4 f4 (36...♔f4 37 a5 ♔f3 38 a6 ♖e2+ 39 ♔d1 ♔xf2 40 ♖f6 ♖xb2 41 ♖xf5+ ♔e3 42 ♖a5 ♔d3 43 ♔c1 ♖b8 44 a7 ♖a8 45 ♖a3 e3 46 ♔d1 e2+ 47 ♔e1 wins for White.) 37 a5 f3 38 a6 ♖e2+ 39 ♔d1! e3 (39...♖xf2 40 a7 ♖f1+ 41 ♔d2 ♖a1 42 ♖b7 ♔f6 [42...f2 43 ♖f7] 43 ♔e3 and White wins the rook ending.) 43 fxe3 ♖xe3 44 a7 f2 45 h4+ ♔f5 46 a8♕ f1♕+ 47 ♔c2 ♖e2+ 48 ♔b3 ♖xb2+ 49 ♔xb2 ♕f2+ 50 ♔c1 ♕xb6 51 ♕f3+ and White has good winning chances in the queen ending. 36...♔g4 37 h3+ White wants to avoid ...f4 and ...♔f3. 37...♔xh3 38 ♖xf5 ♖xb2 The rook ending is now a forced draw. 39 ♖f4 ♖xa2 40 ♖xe4 h5 41 c4 ♔g2 42 ♖f4 ♖c2 43 ♖h4 ♔f3 44 ♔d1 ♖xf2 45 c5 ♔e3 46 ♖xh5 ♔d4 and a draw was agreed. Rubinstein managed to draw by playing as actively as possible on the kingside. Tarrasch had a win but it required exact calculation. This example is reminiscent of how Alekhine managed to draw against Kashdan, position 255. 258 Capablanca – Tartakower New York 1924 White to move This classic position is a world famous example demonstrating in the most instructive manner that activity is more important than material in rook endings. White’s rook and king are much stronger than their counterparts since the black king is cut off on the last rank. If White activates his king on g6 or f6 White will win, since the cooperation between the king, rook and g5-pawn will be too much for Black to handle in the position. Capablanca played the delicious move 35 ♔g3!! The point of the manoeuvre is to walk into the enemy camp on the kingside. 35...♖xc3+ 36 ♔h4 36...♖f3? 36...♖c1? loses quickly after 37 ♔h5! (Not 37 g6? ♖h1+ 38 ♔g5 ♖xh7 39 gxh7 ♔g7 40 ♔xf5 c5! and Black obtains a passed pawn which will secure the draw.) 37...♖h1+ 48 ♔g6 and Black’s f-pawn is lost and White will have two connected passed pawns with an easy win. The best defence was 36...a6! after which Capablanca would have been forced to find 37 ♔h5 b5 38 ♔g6! ♖c6+ 39 ♔xf5 bxa4 (39...b4 40 a5!) 40 ♖h3 with good winning chances. 37 g6! ♖xf4+ 38 ♔g5 ♖e4 39 ♔f6! It’s more important to get the king to the sixth rank than to capture the f-pawn. Note the perfect harmony between the white pieces. The two pawns deficit doesn’t matter when compared to the full activity and cooperation White enjoys on the kingside. 39...♔g8 40 ♖g7+ ♔h8 41 ♖xc7 ♖e8 42 ♔xf5 This is a good moment to capture the f-pawn, as all the black pieces are passive. 42...♖e4 43 ♔f6 ♖f4+ 44 ♔e5 ♖g4 45 g7+ ♔g8 45...♖xg7 46 ♖xg7 ♔xg7 47 ♔xd5 is a simple win. 46 ♖xa7 ♖g1 47 ♔xd5 ♖c1 48 ♔d6 ♖c2 49 d5 ♖c1 50 ♖c7 ♖a1 51 ♔c6 ♖xa4 52 d6 Black resigned. This flawlessly played ending is one of the reasons Capablanca has the reputation of being one of the strongest rook endgame players. However, he had studied more than 1,000 rook endings to achieve this mastery. 259 Korchnoi – Karpov World Championship match, Baguio City 1978 White to move White’s king is more active than its counterpart and can support the breakthrough a5-a6 (or d4-d5) after ♔c3 and ♔b4 and at the same time penetrate the black position via a5 and b6 with pressure on the c6-pawn. Another advantage in White’s position is that he has three pawns on the fifth rank, which is an undoubted advantage if it all comes down to a pawn race. The pawn on f5 provides the white rook with an outpost on e6 which might come in be handy in view of the weak pawn on c6 after White has played a5-a6. Korchnoi played 49 ♔c3 but two stronger possibilites were: a) 49 ♖e6! ♖a8 50 ♖d6 ♖xa5 51 ♖d7+ ♔e8 52 ♖xb7 ♖a3+ 53 ♔e4 ♖xh3 54 ♖c7 ♖h4+ 55 ♔e3 ♖h3+ 56 ♔d2 and Black’s c6pawn falls. White’s two connected passed pawns are more dangeorus than Black’s pawns. b) 49 ♖a2!? ♖g3+ 50. ♔e4 ♖xh3 51 a6 bxa6 52 ♖xa6 ♖h4+ 53 ♔e3 ♖h3+ 54 ♔d2 and Black’s c6-pawn is doomed. 49...♖e8 A more active and better defence was offered by 49...♖g5 50 ♔b4 ♖xf5 51 a6 bxa6 52 ♔a5 ♖d5 53 ♔xa6 (53 ♖e4? f5) 53...♖xd4 54 ♔b6 b4 and Black has enough counterplay to achieve a draw. 50 ♖d2 50 ♖xe8? ♔xe8 leads nowhere since Black has a passed pawn on b5. The pawn ending is a draw. Another possibility was 50 ♖a2!? followed by a5-a6. 50...♖e4 51 ♔b4?! More consequent would have been the breakthrough 51 d5!. Otherwise White could have returned to a plan connected with the breakthrough a5-a6. This could be achieved either with 51 ♖a2 or with an immediate 51 a6 bxa6 52 ♖a2. 51...♔e8 52 a6! bxa6 53 ♔a5 ♔d7 54 ♔b6 It’s more important to eliminate the c6-pawn than the a6-pawn. 54...b4 55 d5! cxd5 56 ♖xd5+ ♔c8 It’s logical to try to block the c-pawn with the king but 56...♔e7 was a draw as well. 57 ♖d3 White must stop the b-pawn from advancing. 57...a5 58 ♖g3 58 ♔xa5 ♖c4 59 ♔b6 ♖c3. 58...b3? Another mistake would be 58...a4? as can be seen after 59 c6 ♖e8 60 ♖g7 b3 61 ♖a7 ♔d8 (or 61...♔b8 62 c7+ ♔c8 63 ♖a8+ ♔d7 64 ♖xe8 ♔xe8 65 c8♕+) 62 ♖a8+ ♔e7 63 ♖xe8+ ♔xe8 64 c7 ♔d7 65 ♔b7 b2 66 c8♕+. The last drawing chance was 58...♖d4! 59 ♔xa5 (59 c6 b3 60 ♖g8+ ♖d8) 59...♔c7 60 ♔b5 ♖d7. 59 ♔c6! 59 ♖xb3?? ♖b4+ 60 ♖xb4 axb4 and Black wins! 59...♔b8 60 ♖xb3+ All of Black’s pawns will be massacred. 60...♔a7 61 ♖b7+ ♔a6 62 ♖b6+ ♔a7 63 ♔b5 a4 64 ♖xf6 ♖f4 65 ♖xh6 a3 66 ♖a6+ ♔b8 67 ♖xa3 ♖xf5 68 ♖g3 ♖f6 69 ♖g8+ ♔c7 70 ♖g7+ ♔c8 71 ♖h7 White resigned. This instructive rook ending showed that it’s difficult even for the best players in the world to realise the right moment for the breakthrough. One should consider that during this period games were adjourned and players also had more thinking time for the last phase of the game. Moreover Korchnoi had the reputation of being a good rook endgame player. He even wrote a book about it, Practical Rook Endings. Don’t forget to study those important rook endings you must know by heart. Spassky used to say that is good advice but to follow it is the harder part! It’s because of such statements that even strong players make mistakes in rook endings again and again. I know several grandmasters who openly admit they ignore the study of rook endings, but the very best players in the world don’t! We now go on to multi-piece endings, with at least two pieces on each side. 260 Centurini Black to move It’s not so difficult for the defender to hold the draw against a rook and knight. The trick is to avoid the opposition and place the rook as far away as possible, while pinning the knight. Here Black keeps the draw by avoiding the opposition with 1...♔c8!. 1...♖c1? loses as shown by Centurini. The key to winning this position is to force the black rook to a worse square. 2 ♖d3! ♖c2 3 ♖d1! Black is in zugzwang and must place the rook on the awkward c4 square which makes it more vulnerable to knight forks. 3...♖c4 4 ♖h1 ♖c2 (4...♔c8? 5 ♘d6+) 5 ♘d4! ♖b2+ (5...♖c4 6 ♖h8+ ♖c8 7 ♘c6+ ♔a8 8 ♖xc8 mate.) 6 ♔c6 ♔a8 (6...♖b4 7 ♖h8+ ♔a7 8 ♘b5+) 7 ♖h3 ♖b1 8 ♖h2 Black is in zugzwang again. 8...♖b4 9 ♘b5 ♖c4+ 10 ♔b6 ♔b8 (10...♖c8 11 ♘c7+ ♔b8 12 ♘a6+ ♔a8 13 ♖a2! [13 ♖h7? fails to 13...♖c6+!.] and Black cannot avoid the double check on c7 next move.) 11 ♘d6 ♖b4+ 12 ♔c6 ♔a8 13 ♖h8+ ♖b8 14 ♘c8! A beautiful end to a beautiful study! 2 ♔c6 ♖c1+ 3 ♔b6 ♖b1 etc... Black easily holds the draw by alternating checks with a pinning of the knight. Don’t forget that even if the defender’s king is driven to the last rank it’s still an easy draw if you avoid the opposition and have an active rook. Try to play this ending with a computer and you’ll see it’s not so difficult to grasp its main principles. 261 Philidor 1749 White to move This is a forced win as was discovered by Philidor way back in 1749. Nevertheless there are many strong players today who fail to win this ending. The correct solution is 1 ♖f8+! First White forces the black rook to the last rank. Note that 1 ♗c6? fails to the typical finesse 1...♖d7+!. 1...♖e8 2 ♖f7 White has conquered the seventh rank and threatens to win at once with ♖a7 followed by ♖a8 mate. 2...♖e2! This is the best defence. 2...♔c8 loses to 3 ♖a7 ♖d8+ 4 ♔c6 ♔b8 5 ♖b7+ ♔a8 (5...♔c8 6 ♗e6+) 6 ♖b5! ♔a7 7 ♖a5+ ♔b8 8 ♔b6 and 8...♔c8 puts the king in a mating net after 9 ♗c6. 3 ♖h7! An important tempo is gained by forcing the black rook away from the second rank. 3...♖e1 The toughest defence. 3...♖e3 loses more quickly after 4 ♖d7+. When Black’s rook is placed on the third rank it’s the signal to force the king to declare its intentions. 4...♔e8 (4...♔c8 5 ♖a7 wins at once since the black rook cannot be transferred to b3 and this is the reason the rook is badly placed on the third rank.) 5 ♖a7 ♔f8 6 ♖f7+ ♔e8 7 ♖f4 (with the threat 8 ♗c6+) 7...♔d8 (7...♖d3 8 ♖g4) 8 ♗e4! This is the point. Black has no check on d3. 8...♔e8 9 ♗c6+. 4 ♖b7! This alternation between both wings is typical for this type of ending. 4...♖c1 4...♔c8 is answered by 5 ♖a7 ♖b1 6 ♖f7 and White wins the rook by force after 6...♔b8 7 ♖f8+ ♔a7 8 ♖a8+ ♔b6 9 ♖b8+. 5 ♗b3! This is the key move. The defender’s rook is forced to enter the fatal third rank. 5...♖c3 5...♔c8 is met by the effective move 6 ♖b4! ♔d8. The idea is to swing it over to the other side of the board. 7 ♖h4 ♖e1 (7...♔c8 8 ♗d5 ♔b8 9 ♖a4) 8 ♗a4! A very beautiful bishop manoeuvre, controlling d1 as well as e8, and a key position to remember. 8...♔c8 9 ♗c6 ♖d1+ 10 ♗d5 ♔b8 11 ♖a4. Black must give up his rook to prevent mate next move. 6 ♗e6 ♖d3+ 7 ♗d5 ♖c3 8 ♖d7+ When the opponent’s rook is placed on the third rank it is important to remember this check to force the king to decide which side to go to. 8...♔c8 8...♔e8 9 ♖g7 wins immediately since ...♖f3 isn’t playable. 9 ♖h7 ♔b8 10 ♖b7+ ♔c8 11 ♖b4! ♔d8 11...♖d3 12 ♖a4. 12 ♗c4! This is the reason the rook must be placed on b4. 12...♔c8 13 ♗e6+ ♔d8 14 ♖b8+ and mate next move. Play through this endgame many times and learn it by heart until you have it at your fingertips. It’s also useful to practice it against a computer. I can guarantee that this endgame will arise more than once in your chess games, both as a defender and as an attacker – so don’t ignore a deep study of it! 262 Timman – Lutz Netherlands 1995 Black to move To be able to draw the defender should be familiar with the second-rank-defence and Cochrane’s method. 79...♖a4+ 80 ♗d4 ♔g5 81 ♖g7+ ♔h4 This is the position we are referring to as Cochrane´s method and it’s a solid defence to play for the draw. 81...♔h5? is a mistake due to 82 ♔f5 leading to the previous Philidor position after 82...♖a5+ 83 ♗e5 when White wins by force. However, 81...♔h6 is playable. 82 ♔e5 If White prepares ♔f4 by defending the bishop with 82 ♖d7 then Black plays 82...♔g4!. If White waits with 82 ♖g8 Black waits as well with 82...♖b4!. 82 ♔e3 is met by 82...♔h5!. 82...♔h3! This is the main idea when adopting Cochrane’s defence. Black’s king goes in the opposite direction to that of the white king. 83 ♖g1 ♖b4 84 ♗e3 84...♖g4! This is the so called second-rankdefence where the cooperation of the white pieces is disturbed. Black could have continued passively with Cochrane’s method after 84...♖a4. It’s just a matter of taste or habit whether you play this or that defence but you should know both of them. 85 ♖a1 ♔g2 Black’s king naturally steps out from the last rank. This is the main point: conquering the second rank. 86 ♗f4 ♖g8 87 ♖a2+ ♔f3 88 ♖a3+ ♔e2 89 ♔e4 ♖e8+ 90 ♗e5 ♖e7 91 ♖a2+ ♔e1 A new Cochrane position has arisen after the position has been rotated 90 degrees. 92 ♔d4 ♔f1! 93 ♗f4 ♖e2! Black is combining the Cochrane method and the secondrankdefence. 94 ♖a8 ♖e7 95 ♔d3 ♔g2 96 ♖f8 ♖e6 97 ♖f7 ♖e8 98 ♗e3 ♖a8 99 ♗c5 ♖a4 100 ♔e3 ♖g4 Black continues the second-rankdefence. If Black wanted to continue according to Cochrane it was possible to play 100...♔g3 101 ♖g7+ ♔h4 102 ♗d4 ♔h5 etc. 101 ♗d6 ♖g6 102 ♖f2+ ♔h3 103 ♗e5 ♔g4 104 ♔e4 ♔h5 105 ♗f6 ♔g4 The main idea with the second-rank-defence is not to allow the king to be driven back to the last rank so if White now checks the king it’s not possible to keep the rook on the second rank since Black’s king slips out again on the next move. 106 ♖f4+ ♔g3 107 ♔e3 ♔h3 108 ♖f5 ♖g3+ 109 ♔f2 ♖g2+ 110 ♔f1 ♖c2 111 ♖g5 ♖c4 111...♖g2! was the most straightforward defence. If the rook is taken it’s stalemate. 112 ♗e5 ♔h4 113 ♖g8 ♖e4 114 ♗g3+ ♔h5 115 ♔f2 ♖a4 115...♖g4! was simpler. 116 ♔f3 ♔h6 117 ♗e5 ♖b4 118 ♗f4+ ♔h7 119 ♖g5 ♖a4 120 ♔g4 ♖b4 121 ♔f5 121...♖b5+ and here the players agreed to a draw. A clever way to play the second-rank-defence is to exploit stalemate possibilities. 121...♖b6 122 ♗e5 ♖g6! 123 ♖h5+ ♖h6 124 ♖g5 ♖g6! is an instructive draw to remember since White cannot avoid the exchange of rooks. 263 Capablanca/Lasker Vossische Zeitung 1914 White to move Capablanca and Lasker met in Berlin 1914 where they played a rapid-transit match of ten games which was won easily by Capablanca. Afterwards a position from one of the games suggested the idea of this study. The solution is 1 ♘xc7 ♘xc7 2 ♖a8+!! A brilliant transition to a winning pawn ending. 2...♘xa8 2...♔xa8 3 ♔xc7 ♔a7 4 ♔c6 obviously makes no difference. 3 ♔c8 Black is in zugzwang. 3...♘c7 4 ♔xc7 ♔a8 5 ♔xb6 ♔b8 264 White to move When I teach young kids pawn endings this position is one of them. But I used to place the kings on h6 and h8 and pawns on b5 and b6 which eventually leads to the diagram position after ♔h6g6-f6 etc. 6 ♔a6! It’s a very common mistake to place the king in the large central area but here it’s an inaccuracy. 6 ♔c6?! is answered by 6...♔a7! when White is forced to repeat the position after 7 ♔c7 (7 b6+?? ♔a8! leads to a draw and is a well-known trap in this ending with a knight pawn.) 7...♔a8 8 ♔b6 ♔b8 9 ♔a6! and White is back on track to win. 6...♔a8 7 b6 ♔b8 8 b7 ♔c7 9 ♔a7 and the pawn is promoted. 265 Dautov – Korchnoi Switzerland 2003 Black to move Rook and knight versus rook and knight is the third to fourth most common endgame featuring two pieces each. According to Flear’s investigations it’s also the most tactical endgame. Small positional advantages like a pawn majority or active pieces become more important with this material distribution. 38...♖a1 Black is a pawn down and has to play a very active game by threatening the b3-pawn or exerting pressure along White’s second rank. 39 ♖d8+ An alternative was 39 ♔e3 ♖c1 (39...♖b1 40 ♖d8+ ♔h7 41 ♘d2) and only now play 40 ♖d8+ ♔h7 41 ♖b8. After 41...♖c3+ 42 ♔d2! ♘d4 43 ♖xb4 ♖xb3 (A better practical chance is perhaps 43...♖xf3 44 ♘e5 ♘xb3+ 45 ♔c2 ♘a1+ 46 ♔b2 ♖f2+ 47 ♔xa1 f6 48 ♖b2! ♖f1+ 49 ♖b1 ♖f2 50 ♘g4 ♖e2 51 ♖b2 ♖xe4 hoping to reach the ending rook versus rook and knight without pawns.) 44 ♖xb3 ♘xb3+ leads to an endgame we have already discussed and where White has good practical chances to win. At a high level the attacker wins more often than draws. 39...♔h7 40 ♖b8 ♘d4 41 ♘d2? 41 ♖xb4? gives Black too much counterplay after 41...♖a2+ 42 ♔f1 ♘xf3. A more active approach was 41 ♘d6 f6 42 ♔e3 ♘xb3 43 ♖xb4 with practical winning chances, although it must be regarded as a success for the defender to be able to hold with three pawns versus four on the same wing. 41...♖a2? Black would have excellent drawing chances after 41...♖h1 42 ♘f1 ♘xf3 and it’s surprising Korchnoi didn’t play like this. Instead the game became more complicated, but in White’s favour. 42 ♔e3 42...♘c2+? The forced variation 42...♖b2 43 ♔xd4 ♖xd2+ 44 ♔e3 ♖xh2 45 ♖xb4 h5 should have been tried. It gives Black reasonable drawing chances due to the active rook behind the passed pawn. Apparently Korchnoi wanted to avoid such a passive continuation in favour of an active defence and calculated risk. 43 ♔e2 White must be careful after 43 ♔d3 ♘e1+ and avoid 44 ♔e2? ♘xf3!. Such tactical finesses are commonplace in this type of ending. 43...♖b2 44 ♖d8! ♘a1 Black cannot utilise the d4-square, which is why he goes to the corner square in order to get maximum activity with pressure on the b3-pawn. 45 ♖d3 f6 46 f4 If White prepares an attack on the black pieces beginning with 46 ♔d1, Black replies 46...♔g6 with the idea that after 47 ♔c1 ♖c2+ 48 ♔b1? Black has the strong reply 48...♖c3!. Such a variation is illuminating for the kind of tactics abundant in endgames where rooks and knights are involved. 46...h5? Correct was the normal 46...♔g8 followed by ...♔f7. 47 ♔d1? Dautov missed a good chance to win after 47 e5! fxe5 48 fxe5 ♔g6 49 ♖f3! (White cuts off the king) 49...♖c2 (49...♘c2 50 ♖f4 ♖a2 51 e6 ♖a7 52 ♔d3 and Black’s position collapses.) 50 e6 ♖c7 51 ♖f4 ♖e7 (51...♘c2 52 ♘f3) 52 ♖xb4 ♖xe6+ 53 ♔d3 etc. 47...♔g6 48 h4? More critical was still the advance of the e-pawn: 48 e5 ♔f5 (48...fxe5 49 ♔c1 ♖c2+ 50 ♔b1 ♖c3 51 ♖d6+ shows the main point of playing the pawn to e5 at move 48.) 49 ♔c1! ♖c2+ 50 ♔b1 ♖c3 51 ♖d5 ♘c2 (51...♘xb3 52 exf6+ ♔xf6 53 ♘e4+ shows tactics again!) 52 exf6+ ♔e6 53 ♖d6+ (53 fxg7? ♘a3+ 54 ♔b2 ♖c2+ 55 ♔a1 ♖c1+ 56 ♘b1 ♖xb1+ 57 ♔a2 ♔f7 58 ♖d4 ♖a1+! 59 ♔b2 ♖b1+ An incredible drawing line revealing the mutual tactical chances inherent in the position.) 53...♔f5 (53...♔xd6 54 ♘e4+ ♔e6 55 ♘xc3 ♘a3+ 56 ♔b2 ♔xf6 57 ♘d5+) 54 f7! (54 fxg7? ♘a3+ 55 ♔a1 ♖c1+ 56 ♘b1 ♖xb1+ 57 ♔a2 ♖c1 58 g8♕ ♖c2+ and Black draws by perpetual check.) 54...♘a3+ 55 ♔a1 ♖c1+ 56 ♘b1 ♖xb1+ 57 ♔a2 ♖c1 58 f8♕+ and White queens with check. 48...♔f7 49 ♖d7+ ♔e8 50 ♖xg7 ♘xb3 51 ♘c4 ♖g2 52 ♘e3 52 ♖b7 ♖xg3 53 ♘d6+ ♔d8 (53...♔f8?? 54 ♖b8+ wins the rook after the following fork.) 54 ♖xb4 ♘c5 and the game should end in a draw due to White’s weak pawns on the kingside. 52...♔f8 53 ♘xg2 53 ♖b7 ♖xg3 is a draw as well. 53...♔xg7 The knight ending is an easy draw since White cannot defend the e4-pawn due to the bad positions of the king and knight. 54 e5 ♘d4 55 exf6+ ♔xf6 56 ♘e3 b3 57 ♔c1 ♘e2+ 58 ♔b2 ♘xg3 59 ♔xb3 ♘e2 60 ♔c4 Here the players agreed to a draw. It was by very active defence that Korchnoi managed to hold the draw and his knight manoeuvre to the corner was indeed impressive. The lesson to be learned is that whenever you reach this type of ending try to see all the tactics in the position. Such endings are therefore especially difficult to play with little time on the clock. 266 Lékó – Adams Dortmund 1996 Black to move This is an instructive position to show that if both players have a pawn majority it’s vital to have greater activity to make the majority a dangerous asset. 32...♖d1+ 33 ♖e1 ♖d2 Normally Black would avoid the exchange of rooks if his rook is more active but in this position it might pay to examine what happens after the exchange: 33...♖xe1+ 34 ♔xe1 e4!. This is the move – making it worthwhile to analyse a little bit further. White cannot avoid a serious weakening of the f4-square since 35 ♔f2 exf3 36 gxf3 ♔d6 37 ♔e3 ♔e5 is literally forced. However, White has to be careful. If he plays 38 ♗e2 followed by c2-c3 and b3-b4 he should be able to draw. However, the seemingly innocent 38 c3? loses the pawn ending after 38...♗d5! 39 ♗xd5 ♔xd5. 267 White to move The reader who doesn’t believe or understand why this pawn ending is lost should look at it in more detail. Black’s plan is simply to play ...b6-b5 followed by ...a5-a4. In such a situation Black’s threats are twofold: a) He wants to attack the c-pawn with the king. b) He wants to create a passed pawn on the kingside and lure the defending king to that side of the board. In the meantime Black picks up the c- and a-pawn and then queens the a-pawn. There is not much White can do. Play might go 40 c4+ ♔e5! 41 b4 f5 42 h3 h4! 43 c5 bxc5 44 bxa5 f4+ 45 ♔d3 ♔d6 and White is unable to defend the a-pawn or attack the c-pawn due to the decisive break ...g5-g4. 34 ♖e2 ♖d4 35 ♔f2 35 h3 would have slowed down Black’s advance of the kingside pawns. However, the drawback is the dark-square weaknesses on the kingside which might tell in the future. 35...g4 In this position it’s clearly apparent that the black pawn majority, combined with Black’s active rook and bishop (and king!), is more dangerous than White’s which hasn’t yet begun to advance. 36 ♗d3 h4!? Black plans to disrupt White’s pawns on the long diagonal with ...h4-h3. The normal continuation was the activation of the king. After 36...♔e6, preparing ...f5, the bishop checks 37 ♗c4+ ♔f5 38 ♗d3+ ♔g5 are not disturbing as they help Black to place his king on the same side as the majority. Black’s idea is to establish his king on f4 after a preliminary ...gxf3. 37 fxg4?! The drawback is obviously that Black now has a dangerous passed pawn on the e-file. A more clever defence was 37 ♖e3!, planning to meet 37...h3 by 38 c3 ♖d8 39 fxg4 hxg2 40 h4 with counterplay. 37...♖f4+ 38 ♔g1 ♖xg4 39 c3?! It’s easy to understand White’s reluctance to mobilise his majority on the queenside, especially now when Black has already established a passer. However, the problem is that the pawns will not travel far due to the weakness of the c4-square after the pawn advances to b4. From that point of view it’s better to play a waiting game with moves like 39 ♔f1 or 39 ♖d2. 39...♗d5! 40 b4 ♔e6 Black now has a very strong centralised position and prepares ...f6-f5. 41 ♖f2 41 ♗a6 f5 42 ♖f2 ♖g8 makes no real difference 41...♖g5 A more effective continuation was 41...e4! 42 ♗e2 ♖g7, not worrying about the mobilisation of the pawn majority since Black’s position is more active and centralised. An illuminating variation would be the following: 43 c4 ♗c6 44 ♗f1 axb4 45 axb4 f5 46 ♖f4 e3! 47 ♖xh4 ♔e5 48 ♖h6 ♗e4 49 ♖xb6 f4 and Black wins since ...f4-f3 is coming. 42 c4 ♗c6 43 ♗f1 axb4 44 axb4 f5 45 ♖a2 f4 45...e4 was also good. 46 ♖a6 ♗xg2! Time to harvest! 47 ♖xb6+ ♔f5 48 ♖b8 h3 49 c5 e4 50 ♖f8+ 50 c6 e3 51 c7 leads to checkmate in four: 51...♗e4+ 52 ♗g2 ♖xg2+ 53 ♔f1 ♗d3+ 54 ♔e1 ♖g1 mate.. 50...♔e5 51 c6 e3 52 ♗xg2 ♖xg2+ 53 ♔f1 ♖c2 54 b5 ♔e4 Both pawn majorities have advanced far but the deciding factors are Black’s active king and the control of White’s second rank. 55 ♔e1 ♔f3 Philidor would presumably have played 55...f3 and that’s the best move – with a forced mate in 14 moves. 56 ♔d1 ♖c5 57 ♖a8 ♔f2 Black plays according to the no-hurry-principle. 57...e2+ was simpler as well as faster. 268 Bauer – Flear France 2006 Black to move From his practical experience Flear thinks that it’s not easy to win this type of endgame despite the fact that White has a distant passed pawn. It’s useful to remember that it pays to exploit all the defensive possibilities in the position. 41...♔f8! Centralisation with 41...♔f6 means problems with the f7-pawn after 42 ♖a7, threatening ♗e8. 42 ♖a7 g5 Black gives the king some air while planning to fix the f2-pawn with ...g5-g4. 43 hxg5 White doesn’t mind exchanging his worst pawn but the brutal 43 a5! deserved serious attention. Then after 43...g4 (43...gxh4 44 gxh4 ♘d3 4 a6 ♘e1+ 46 ♔f1 ♘d3 47 ♔e2 ♘f4+ 48 ♔d2 ♘e6 49 ♗c4!) 44 a6 ♘d3 45 ♖a8+ ♔g7 46 ♖e8! ♘b4 (46...♘c5 47 ♖c8! ♘e6 48 ♖c4) 47 ♖e7! ♔f6 48 ♖c7 ♘d5 49 ♖c6+ ♔g7 50 ♗c4 ♘c3 51 ♖c7 White wins a second pawn. Note the perfect cooperation displayed by the rook and bishop in all these variations. 43...hxg5 44 a5 ♘d3 45 ♔f3 Flear believes that it’s doubtful whether White can win the endgame after 45 ♗xd3 ♖xd3 46 ♖b7 ♖a3 47 ♖b5 f6. As usual, the attacker must be careful before making an exchange when playing with two pieces against two pieces. You are supposed to exchange only when it’s certain it leads to an advantageous endgame, otherwise you obviously avoid the exchange. White should not worry too much about the king and rather strengthen the situation on the queenside. The strongest continuation was 45 a6! ♘e5 46 ♖b7 g4 47 a7 ♔g7 48 ♔f1 ♘f3 49 ♗d7 (with the idea ♔e2, ♔d1, etc.) 49...♘e5 50 ♗e6 followed by ♗d5 with a win. 45...♘c5+ 46 ♔g4 ♘e6 47 ♗d7 ♘c5 48 ♗c6 ♘d3 49 ♗d5 ♘xf2+ 50 ♔f5 White has succeeded in obtaining a very active position so Black must defend precisely if he should have any chance to survive. 50 ♔xg5 is obviously met by 50...♖xg3+ 50...♖d3! 51 ♖xf7+ ♔e8 52 ♔e6 Fritz recommends the beautiful 52 ♖h7! ♖xg3 (52...♖xd5+? 53 ♔e6) 53 ♔e6 ♖e3+ 54 ♔d6 ♘e4+ 55 ♗xe4 ♖xe4 56 a6 with a win. 52...♖e3+ 52...♖xd5? 53 ♔xd5 ♔xf7 54 a6. 53 ♔f6 Another winning continuation was 53 ♔d6 ♘e4+ 54 ♔c6 ♖a3 (54...♘xg3 is answered by 55 ♖f3!.) 55 ♔b6 ♘d6 56 ♖h7. 53...♘g4+ 54 ♔g7 ♖a3 55 ♖f5? Correct was 55 ♖a7 ♔d8 56 ♗e6 ♘f2 57 ♗f5 and White wins. 55...♔e7 56 ♖xg5 ♖xg3! 56...♖xa5? 57 ♖xg4 ♖xd5 58 ♖e4+ and White wins the rook ending by setting up a Lucena position. 57 a6 ♖a3 58 ♖xg4 ♖xa6 The famous endgame rook versus rook and bishop without pawns has arisen and we have already discussed the way to win and to draw in positions 261-262. Don’t forget to learn the winning and defensive methods since endings where rook and bishop versus rook are involved, with pawns as well as without pawns, will arise in every 63rd game! If you know this particular ending without pawns by heart, then due to the resources present there, it’s possible to save many difficult endings where there are pawns involved. In other words: even if the ending doesn’t arise on the board it certainly does when you calculate the variations. 59 ♖e4+ ♔d7 60 ♔f7 ♖a1 61 ♗e6+ ♔d6 62 ♔f6 ♖d1 63 ♗g4 ♖f1+ 64 ♗f5 ♖d1 65 ♖e6+ ♔c5 66 ♗e4?! 66 ♖e8 retains some practical winning chances but of course the position is a draw if Black plays in accordance with the rather easy second rank defence. 66...♖d6 and a draw was agreed. The lesson we can learn from this endgame is that although White had an active rook and bishop as well as an extra pawn, which was an outside passer, it demanded precise play to win. It’s difficult to win on the queenside alone so White needed an additional weakness on the kingside. With rook and knight and perhaps one pawn, Black can create annoying tactical possibilites and make it harder for White to win. Black actually laid the foundations by exchanging pawns on the kingside and then managed to survive a piece down after all the pawns had been exchanged. 269 Fischer – Taimanov Vancouver 1971 White to move Rook and bishop versus rook and knight is the most common endgame with two pieces each and it will arise approximately in every eleventh game. Such statistical information means it’s necessary to make a special study of this ending and learn the principles of playing with this material relationship. According to the general rule, rook and bishop are stronger than rook and knight but Flear doesn’t agree with this superficial estimation since it doesn’t tally with his careful investigation. Of course we receive this important knowledge with a warm heart and become aware of it whenever we are in contact with this type of ending. Primarily, very small changes in the pawn structure or a more active king can change the general evaluation of such a position drastically. 25 ♗f1 Here the bishop is definitely stronger than the knight since there is a completely open centre and possibilities for playing on both wings. Now the bishop puts pressure on the f1-a6 diagonal and the a6-pawn is forced to declare its intentions. The knight lacks an active outpost and therefore this is a difficult ending to handle for Black. When playing with two pieces it’s very important that both of them are gainfully employed because if one of them is ineffective it can be enough to lose the game because of that factor alone. The ideal situation for the defender would be to have his knight placed on d4 and the c2-pawn on c4. This change would not only improve the knight dramatically, it would also make the bishop significantly worse. The knight usually prefers closed positions or those with a shorter front line. 25...a5 26 ♗c4 ♖f8 Black’s pieces are considerably more passively placed than White’s. However, Black has a more active king. 27 ♔g2 As a rule, the king should remain outside the knight’s radius of action to avoid an unpleasant fork. 27...♔d6 28 ♔f3 Now White’s king is active as well. 28...♘d7 29 ♖e3 ♘b8 In this position it’s difficult to play for a liberation of the rook by placing the f-pawn on f6, due to the weakness of the e6-square. 30 ♖d3+ ♔c7 31 c3 A typical prophylactic move to take the sting out of the knight by protecting the d4-square. Note that the f4-pawn also controls an important square in the centre which is not available for the black knight placed on c6. 31...♘c6 32 ♖e3 ♔d6 33 a4 It’s correct to break Capablanca’s famous rule and place this pawn on a white square, since there is no way for Black to attack it. It’s more important to fix the b6-pawn and then create possible squares for a future incursion of the king via b5. 33...♘e7 34 h3 ♘c6 35 h4 White plays according to the no-hurry-principle and this is an important psychological rule when in control of the position to make your opponent more tired before concrete variations and concrete breakthroughs take over the game. 35...h5?! Black breaks the famous rule by Capablanca as well, but in this case the h5-pawn is a future target for the bishop or the rook. It was better to play a waiting game such as by 35...♘e7. 36 ♖d3+ ♔c7 37 ♖d5 Now White is forced to place another pawn on a light square which only makes the bishop more than happy. 37...f5 38 ♖d2 ♖f6 39 ♖e2 ♔d7 40 ♖e3 g6 This is unavoidable in the long run since White can easily provoke these weaknesses anyway. 41 ♗b5 ♖d6 42 ♔e2 ♔d8?! 42...♔c7 43 ♖d3 ♖e6+ was better to avoid the exchange of rooks. 43 ♖d3! This is absolutely the right moment to exchange into an advantageous ending with bishop versus knight since Black has pronounced weaknesses on both sides of the board. 43...♔c7 44 ♖xd6 ♔xd6 270 Fischer – Taimanov Vancouver 1971 White to move 45 ♔d3 ♘e7 46 ♗e8 ♔d5 47 ♗f7+ The bishop not only hits the g6-weakness but also forces Black to decentralise his king. Note the nice cooperation between the pawns, controlling the centre, and the bishop, controlling the light squares in the centre. 47...♔d6 48 ♔c4 ♔c6 49 ♗e8+ Again White uses the power inherent in the bishop to displace the black king. 49...♔b7 50 ♔b5 ♘c8 51 ♗c6+ An important check. 51 ♗xg6?? ♘d6 mate shows how careful White must be of such mates! 51...♔c7 52 ♗d5 ♘e7 53 ♗f7 ♔b7 54 ♗b3 ♔a7 55 ♗d1! ♔b7 56 ♗f3+ We have just witnessed a nice manoeuvre to the long diagonal. We will see the importance of the long diagonal(s) when we come to the queen endings, positions 294-295. The importance of the long diagonal(s) is always something to think about whether we have bishops or queens on the board. 56...♔c7 57 ♔a6 ♘c8 58 ♗d5 ♘e7 59 ♗c4 ♘c6 60 ♗f7 ♘e7 61 ♗e8 It’s typical for this kind of ending that one wins with bishop versus knight when playing according to the principle of the two weaknesses. Black is in zugzwang. 61...♔d8 The only move to avoid immediate loss of material but now White exploits his more active position with the following beautiful and decisive move. 62 ♗xg6! ♘xg6 63 ♔xb6 ♔d7 64 ♔xc5 White has three pawns for a knight and still keeps the more active position. 64...♘e7 65 b4! This is actually the only move to win so White has the time on his side as well. 65...axb4 66 cxb4 ♘c8 67 a5 ♘d6 68 b5 ♘e4+ 69 ♔b6 ♔c8 70 ♔c6 ♔b8 71 b6 and Taimanov resigned. Don’t ever forget this classic! And, above all, replay it on a regular basis! Fischer’s favourite piece was the light-squared bishop. You should ask yourself: Which piece is yours? 271 Bologan – Salov France 1999 White to move This is the third to fourth most common endgame with two pieces on each side. It will occur in every 25th game that you play. What’s important to think about is that small positional advantages can be decisive if the opponent lacks counterplay and is unable to generate threats or activity. This situation might very well happen if you are saddled with a bad bishop. The game continued 61 ♖c5 ♗d7 62 ♗d3 ♔d6 63 ♗c4 ♔c7 64 ♖e5 ♔d6 65 ♗d3 ♔c7 66 ♖c5 ♔d6. White has correctly played according to the no-hurry-principle but how can White play for a win in a more concrete fashion when it’s not possible for his king to enter Black’s position? The answer is that White simply has to open the game to exploit his advantage in mobility. 67 b4! It might seem paradoxical to exchange Black’s weakness on a5 and liberate Black’s rook but this is the kind of creative technique which leads to a situation where the passed pawn will be dangerous and so the principle of the two weaknesses is on the agenda. 67...axb4+ 68 ♔xb4 ♖b8+ 69 ♔c3 ♖a8 70 a5 ♔c7 71 ♗e2 71...♖b8 71...♔d6 can be answered by 72 a6, planning ♔b4-a5, so a logical continuation is 72...♔c7 73 ♖a5! ♔b6 74 ♔b4 ♗c8 75 ♖a3! ♗xa6 (Playing a waiting game with 75...♖a6 is met by the ingenious 76 ♗c4 ♖a8 77 ♖b3!! ♗xa6 78 ♖a3 ♗b7 79 ♖xa8 ♗xa8 80 f5! gxf5 81 ♗e2 and White obtains a passed pawn on the h-file.) 76 ♖xa6+ ♖xa6 77 ♗xa6 ♔xa6. 272 White to move This pawn ending is a good example to show that a king’s activity is sometimes worth more than a pawn. White wins after 78 ♔c5 ♔b7 79 ♔d6 ♔b6 80 ♔e7 ♔b5 81 ♔xf7 ♔c4 82 ♔xe6 ♔xd4 83 ♔f6 c5 84 ♔xg6 c4 85 f5 c3 86 f6 c2 87 f7 c1♕ 88 f8♕. The queen ending is winning after 88... ♕e1 89 ♕d6+ ♔c4 90 ♔xh5. 72 ♖c4 White plans to place the rook behind the passer but allows Black to activate his rook. This is creative technique but perhaps not the best. It’s possible to play 72 ♗d3, for example 72...♖a8 73 a6 ♔b6 74 ♔b4 followed by ♖a5 as mentioned earlier. 72...♖b1 73 ♖a4 ♖g1 73...c5 74 ♖c4 wins the pawn while 73...♔b8? 74 ♖b4+ ♖xb4 75 ♔xb4 leads to a clearly won bishop ending after 75...♗c8 76 ♔c5 ♔c7 77 a6. 74 a6 ♗c8 75 a7 ♗b7 76 a8♕ ♗xa8 77 ♖xa8 ♖xg3+ 78 ♔c4 ♔d7? Black should have continued 78...♔d6, not allowing the king to penetrate to c5. 79 ♔c5? Stronger was 79 ♖a7+ ♔e8 80 f5! and if 80...exf5 81 ♔c5 and ♗c4. 79...♖h3? 79...♔e7! was necessary to be able to meet 80 ♖a7+ with 80...♔f6. 80 ♖a7+ ♔e8 81 f5! An important pawn break to exploit the extra piece. White is winning. 81...♖xh4 81...exf5 is met by 82 ♗c4 and 81...gxf5 by 82 ♗xh5 ♖xh4 83 ♗xf7+ ♔f8 84 ♗xe6. Black’s pawns fall like a line of dominoes and then the d-pawn decides. 82 fxg6? Salov should have played 82 f6!! deciding the game immediately due to the beautiful variation 82...♖f4 83 ♔d6 ♖xd4+ 84 ♔xc6 ♖d8 85 ♗b5 ♔f8 86 ♔c7 when the rook is unable to move since it must cover the last rank. 82...fxg6 83 ♗f3 ♔f8 84 ♗xc6 ♖f4 85 ♖a1 ♔f7 86 ♖e1 ♖f5+ 87 ♔d6 87 ♖e5 was also good enough for a win. 87...♖f4 88 ♗e4 88 ♖e4 was the alternative. 88...♔f6 89 ♔c5 h4 90 ♗c6 h3 91 ♗d7 h2 92 ♖xe6+ ♔g5 93 ♖e1 ♖f2 94 ♗c6 ♔f4 95 d5 ♖c2+ 96 ♔b6 ♖d2 97 ♔c7 g5 98 d6 g4 99 d7 ♔g3 100 d8♕ ♖xd8 101 ♔xd8 ♔f2 102 ♖h1 g3 103 ♔e7 Black resigned. 273 Tukmakov – Wojtkiewicz Bern 1993 White to move Endings with two rooks versus two rooks are the fifth most common ending with two pieces each and they are important to study since they will arise in every 25th game. Here the winning chances are increased compared with one rook versus one rook. White played 35 ♖bb7 reducing one of Black’s rooks to passivity. 35...♖f8 36 g4! This is a standard move to prevent Black from setting up the strong defensive pawn wave formation ...h6-h5 followed by ...g7-g6. 36...♖a4 37 h3 37...h5?! There was no reason to weaken the kingside at this moment and at the same time give White more space on which to operate. 37...♖aa8! was the most logical move, trying to swap a pair of rooks with ...♖ac8, but not ...♖ab8? because of ♖xf7!. Such tricks are typical in double rook endings and one must always watch out for them. White is actually forced to leave the seventh rank with one of his rooks unless he wants to play for a win according to the model of Capablanca – Duras/Yates. After all, White’s winning chances have improved a little due to the important inclusion of g3-g4. 38 g5! 38 gxh5 is answered by 38...♖h4, for example 39 ♖c5 ♖xh3 40 ♖b4 f5! and the rook avoids being trapped. 38...g6 39 ♔g2 ♖aa8 40 ♖c6! This rook has a nice outpost on f6 and at the same time averts the forced exchange after ...♖ac8. 40...♖ab8 41 ♖bc7 ♖b5 42 h4 ♖b4 43 f3! 43...♖fb8 43...♖xh4? 44 e4, threatening ♖c1 and ♔g3, loses after 44...♖a8 45 ♖f6 ♖a5 46 ♔g3 ♖h1 47 ♖cxf7 ♖xg5+ 48 ♔f4 and Black cannot move the rook due to mate in three moves. Such mating constructions are also typical of double rook endings. 44 ♖f6 ♖8b7? It’s vital to keep the f-pawn. Now White’s central pawns easily decide the game. Black’s last chance was 44...♖f8. After 45 e4 ♖b5 46 ♖e7 White’s e5-e6 cannot be prevented. But if Black defends accurately and plays 46...♖b2+ 47 ♔g3 ♖b3! it’s not clear how White can win. 274 White to move This position and this idea is very important to remember. The right location for the active rook is where it puts pressure on the only weakness in White’s position, the f3-pawn. This is important as the sequel will show: 48 e5 ♖a3 49 e6 fxe6 50 ♖xg6+ ♔h8 and now White must take care of the f3-pawn by playing 51 ♖f6. However, Black has a forced draw after 51...♖xf6 52 gxf6 ♔g8 53 ♖xe6 ♔f7 54 ♖e5 ♔xf6 55 ♖xh5 ♖a1 reaching the notorious drawn ending with rook and bishop pawn. If you are familiar with Gligoric – Smyslov it isn’t so difficult to secure the draw. In the game Black had no chance at all after 45 ♖fxf7 ♖xc7 46 ♖xc7 ♖xh4 47 e4 ♔f8 48 ♖c8+ ♔g7 49 ♖c7+ ♔f8 50 ♖c1 ♖f4 51 ♖c8+ ♔g7 52 ♖c7+ ♔g8 53 ♖c6 ♔g7 54 ♖c7+ ♔g8 55 ♔g3 ♖f8 56 f4 ♖e8 57 ♖c4 ♖f8 58 ♖c5 ♖e8 59 ♔f3 ♖a8 60 f5 ♖a3+ 61 ♔f4 h4 62 ♖c8+ ♔g7 63 f6+ ♔f7 64 ♖c7+ ♔f8 65 ♖g7 and Black was forced to resign. 275 Ivanchuk – Short Amsterdam 1994 White to move Ivanchuk played 25 h4!. The weak queenside and the isolated pawn on e5 are not enough for White to win so he must provoke a third weakness. The idea of advancing the h-pawn is to play the g-pawn to g6 with the help of the h-pawn. White must expand on the kingside to be able to win. That’s the only chance. 25...♖e8 Black wants to make a four-rookswap and then centralise the king on e6 and quickly play ...b7-b5 and ...♘b7. This forces White to play b2-b4 after which Black can get rid of his doubled pawn and play ...c6-c5. The natural 25...h5 means that Black hopes to keep the position closed but that strategy isn’t realistic, since White can play a timely g3-g4 and activate his king in the process. 26 g4 White wants to play his pawns to g5 and h5, followed by ♘g4 with pressure on the e5-pawn. 26...♘b5? White pawns on h5 and g4 against g7 and h6 lead to a structure where White can occupy the f5-square with his knight. After an exchange on f5 White will have a pawn majority on the kingside following e4xf5. 27 ♖xf7 ♔xf7 28 ♖f1+ ♔e6 29 ♘f5 ♖g8 30 g5! White has some practical winning chances. His dream plan is to continue ♔g2-g3-g4 followed by h4-h5 and g5-g6 and if Black exchanges on g6 then White continues ♔g5 and ♖f1-h1-h7 with pressure on the g7-pawn. White won after 48 moves. 276 J. Polgar – Lazić Dortmund 1990 White to move Tibor Karolyi has written an excellent book Judit Polgar – The Princess of Chess and in the game Polgar – Lazic he places two exclamation marks after Polgar’s move 22 ♖ad1!!. He regards this as a beautiful move because it’s an original finesse to lure the e7-rook to d7 and in that way manage to defend the d5-pawn indirectly by playing the rook to c1 with pressure on the c7-pawn. This motif is very unusual but good to know, since correct play with two rooks can be tricky with all those ranks and files to think about. The natural move 22 ♖fd1 doesn’t work because of the mating threat on White’s first rank after 22...♖ed7 23 ♖ac1 ♖xd5 24 ♖xd5 ♖xd5 25 ♔f1 c6 and White is a pawn down. 22...♖ed7 If Black rejects placing the d5-pawn under fire by playing 22...♗f8, White will improve the placements of his rooks after 23 ♖d3 ♔g7 24 ♖c1. 23 ♖c1 ♖f7 24 ♖fd1 The net result is that White has gained a tempo with a clever operation of the rooks. 24...♖fd7 25 ♔f1! White’s position is improved with every move. 25...♔h8 26 ♔e2 ♗f8 27 ♖d3 h6 27...♔g7 28 ♘g5 ♔f6 29 ♘e6. 28 ♖b3! White is focusing on the tactical weakness on b7. 28...♖b8 Of course if 28...b6 then 29 ♖c6. 29 ♖c6 bxc6 30 ♖xb8 ♔g7 31 dxc6 ♖d4 32 ♘d2 ♖d6 33 ♘c4 ♖xc6 34 b3 and Polgar converted her advantage to a win after 68 moves. 277 Marin – Engqvist Stockholm 2001 White to move White has just played Kramnik’s suggestion 20 ♔b1 which plans the rook manoeuvre ♖c1-c6 in conjunction with ♖b3. Black played 20...a5 to be able to answer ♖c1 with ...♖ac8 as the a7-pawn is no longer hanging. This is actually Onischuk’s clever idea which also prepares a minority attack to create a weakness on the queenside. 21 a3 ♗e7 22 ♖e1 ♔f7?! Correct was optimal activation of the king by 22...♔f5 followed by ...g7-g5, ...h5-h4 trammelling White’s pawns on the kingside. This idea I used successfully in a later game against IM Dan Cramling. 23 ♖c1 ♖d7 24 ♖c6 a4 White’s b2-pawn is fixed and is now a future weakness on which Black can focus. 25 ♔c2 f5 Marin points out that this move weakens the e5-square and the long dark-squared diagonal, but nevertheless it seems playable. 26 ♗d4 ♗g5 27 ♗e5 ♗h4 28 ♖d2 ♖aa7 29 f4 ♗e7 30 ♖d3 ♖a8!? 30...♗f6 was a more active attempt to play for a draw. 31 ♔d2 ♔g8?! This is not only passive but also planless. The king was better placed on f7. 32 ♖dc3 ♖aa7 33 ♖g3 ♗h4 34 ♖c8+ ♖d8 35 ♖xd8+ ♗xd8 36 ♖d3 ♗b6 37 ♔e2 ♖d7?! It’s important to play actively with the rook, so correct was 37...♖a5 with good drawing chances after 38 ♖c3 (or 38 ♗d4 ♗xd4 39 ♖xd4 ♔f7) 38...♗c5!. 38 ♖c3 d4? The turning point. 38...♔f7 with drawing chances was correct. 39 ♖c8+ ♔f7 40 ♔d3 ♗a7? 40...g6 41 ♖c4 ♖a7 42 ♗xd4 ♗xd4 had to be played. The active a4-pawn has in principle the same value as the queenside pawns. 41 ♖c4 and White won easily. 278-300: Queen Endings A characteristic feature of queen endings is that one must always be aware of the safety of the king, especially when there are more pieces cooperating with the queen such as one or two rooks. They demand a lot of calculation and therefore a study of the most important positions is mandatory in order to make them easier to handle. Among these positions is a high frequency of queen endings without pawns, but that’s because it will then be easier to understand the peculiarities of the queen. The deep knowledge that will be assimilated can then be used in other positions by way of analogy. So it’s not the endgame itself that is important, it’s a knowledge of the queen’s innate capabilities. 278 Philidor 1803 White to move White wins easily against a central pawn or a knight pawn since there are no stalemate resources. 1 ♕f8+ ♔g2 2 ♕b4 A clever move to shorten the staircase. 2...♔f2 3 ♕f4+ ♔g2 4 ♕e3 ♔f1 5 ♕f3+ This is the position to keep in mind. Black cannot go to g1 and is forced to place the king in front of the pawn. This gives White time to improve his king. 5...♔e1 6 ♔b6 ♔d2 7 ♕f2 ♔d1 8 ♕d4+ ♔c2 9 ♕e3 ♔d1 10 ♕d3 White repeats the same pattern again until the king is close enough to deliver a mate. 10...♔e1 11 ♔c5 ♔f2 12 ♕d2 ♔f1 13 ♕f4+ ♔g2 14 ♕e3 ♔f1 15 ♕f3+ ♔e1 16 ♔d4 ♔d2 17 ♕c3+ ♔d1 18 ♔d3 e1♕ 19 ♕c2 or 19 ♕a1 mate. 279 Lolli 1763 Black to move 1...♔c2! or 1...♔c1! is correct. It’s a mistake to place the king in front of the pawn as after 1...♔a1? 2 ♔b6! ♔b1 3 ♔c5+ the white king is close enough to weave a mating net with the queen. For example 3...♔c1 4 ♕a8 ♔b2 5 ♕g2+ ♔b1 6 ♔b4! a1♕ 7 ♔b3. The newly born queen is too awkwardly placed to prevent mate or loss of the queen. 2 ♕a6 ♔b2 3 ♕b5+ ♔c2 4 ♕a4+ ♔b2 5 ♕b4+ ♔c2 6 ♕a3 ♔b1 7 ♕b3+ ♔a1 Now it’s obviously the right moment to step into the corner. White’s king is too far away and the position is a draw. 280 Horowitz 1965 White to move Queen against pawn is normally a win unless there is a rook pawn or a bishop pawn on the seventh rank. If that is the case the attacker’s king must be close enough to be able to win. The study by Horowitz shows how to win when the pawn is on the sixth rank. 1 ♕h6!! Surprisingly this is the only way to win. Beware of the stalemate trap after 1 ♕c2?? or 1 ♕d3?? c2! with an elementary draw. For example 2 ♕a3+ ♔b1 3 ♕b3+ ♔a1!. 1...♔b2 1...♔b1 is answered by 2 ♕b6+ ♔c1 3 ♕e3+ (or 3 ♕g1+) 3...♔b2 4 ♕d4! (or 4 ♕e5!) and the pawn is pinned as in the main variation. 2 ♕f6! This is the point. When the pawn is on the sixth rank it’s important to pin it! 2...♔c2 2...♔b3 is met by 3 ♔g6 when 3...c2 leads to nothing after 4 ♕a1! with an easy win. 3 ♔g6 ♔d2 4 ♕d4+ ♔c2 5 ♔f5 and White wins. 281 Bilguer 1843 White to move When I studied this endgame from Pachman’s Endspiel im Schach I learned that it’s a draw. But this concrete position is actually a loss according to Nalimov’s Tablebase. It’s a forced mate in 43 moves starting with the logical move 1 ♕c7! cutting off the king on the last rank. According to Pachman and Chéron White should play 1 ♕e6? but this is a draw after 1...♔g7. The queen looks strong on e6 and it is. White plays the waiting move 2 ♔f3 and now it’s easy to go wrong according to Tablebase. Pachman/ Chéron analyse 2...♘h7? (Correct is 2...♘g8! with a draw.) 3 ♔g4 ♘hf8 4 ♕d6 ♔f7 5 ♕d5+ ♔g7 6 ♔g5?? Apparently Pachman/Chéron wanted to show the strength of the black knights after 6...♘h7+ when Black wins the queen. The best move is 6 ♕b7+! which shows the real strength of the queen. It’s a forced mate after another 30 moves. 1...♘e7 2 ♕b7 White prepares ♔f4 and steps aside from a possible fork. 2...♘fd5 3 ♔f3 To the centre! 3...♔f7 4 ♕b3 ♔e6 5 ♔e4 ♔d6 6 ♕d3 ♔e6 7 ♕h3+ ♔d6 8 ♕h6+ ♔c5 9 ♕f8 It’s hard to explain the reasons for the queen moves since this is brute force by the computer. 9...♔d6 10 ♔d4 ♔e6 11 ♕h6+ ♘f6 12 ♕e3+ ♔d7 13 ♕d2 ♔d6 14 ♕b4+ ♔e6 15 ♕b6+ ♔f7 16 ♕b3+ ♔e8 17 ♔e5 Through the centre! We can certainly see that White has been successful with the advance of the king which is now approaching the heart of Black’s position. 17...♘d7+ 18 ♔d6 ♔d8 19 ♕d3 ♘f8 20 ♕d1 ♔e8 21 ♕e2 ♘fg6 22 ♕e6 White is now deep in enemy territory with both his pieces. When the knights are in a position where they defend each other, the attacker has been successful since it will be easier to play for a mate. Black’s knights are no longer coordinating well and only step on each other’s toes. 22...♔f8 23 ♔d7 ♔g7 24 ♔e8 White’s king approaches deeper and deeper into the brain of Black’s position. 24...♔h6 25 ♔f7 ♔g5 26 ♕e4 Now it’s zugzwang. 26...♘f5 The best move according to Tablebase. The human move 26...♔h5 is met by 27 ♕g2! ♔h6 28 ♕g3! ♔h5 29 ♕g1!! ♔h4 30 ♔f6 ♔h3 31 ♔g5 and Black is in zugzwang since he has to move one of the knights. 27 ♕g2+ ♔f4 28 ♕xg6 ♘d4 29 ♔f6 The technique for winning against a lone knight is the same as if the king was on its own. White plays for mate rather than winning the knight. 29...♔e3 30 ♕e8+ ♔f3 31 ♕e7 ♘e2 32 ♕e5 ♘g1 33 ♕d4 ♘e2 34 ♕d3+ ♔f2 35 ♔f5 ♘g1 36 ♔e4 ♔g2 37 ♕d1 ♘h3 38 ♕f3+ ♔h2 39 ♔e3 ♘f4 40 ♕f2+ ♔h3 41 ♔f3 ♘h5 42 ♕g1 ♘f6 43 ♕g3 mate. What is really impressive about this endgame is how the attacker uses the queen over practically the whole board. Particularly the waiting moves on the last rank are a key part of the beautiful manoeuvring technique. 282 Pachman Endpielpraxis im Schach 1977 White to move When the knight is protected by the bishop and the bishop is protected by the king the defender has the possibility of constructing a fortress. In the following position Black cannot break through if White is careful to move the bishop from b2 to a1 and back to b2 and so on. 1...♔d3 The black king can penetrate the d-file but it’s of no use. 1...♕d1+ 2 ♔a2 ♕d3 3 ♗a1 ♕d2+ 4 ♗b2 (The placement of the king on b1 or a2 is irrelevant for the result, so 4 ♔b1 is a draw as well.). 2 ♗a1 Note that 2 ♔c1? is a serious error because of 2...♕a2! when White loses by force after 3 ♘f3 ♕c4+ 4 ♔b1 ♕c2+ 5 ♔a2 ♕a4+ 6 ♗a3 (6 ♔b1 ♕d1+ and the knight is lost.) 6...♔c2 7 ♘e1+ ♔c3 and Black is threatening mate in three moves. The lesson to be learned is that the position is lost if the defender lacks harmony in his position. This has been known since 1843 when Bilguer showed a study to prove it. A famous practical example is Pachman – Guimard, 1955. 2...♔d2 3 ♗b2 ♔d1 4 ♗a1 ♕b4+ 5 ♗b2 Even 5 ♔a2, temporarily yielding the c1-square to Black’s king, holds the draw as after 5...♔c1 White has a saving check either on e2 or b3 and then returns with the knight to d4. 5...♔d2 6 ♔a2 ...and Black cannot improve his position. It has been known since 1903, in a study by Karstedt, that this is a drawn position. Black can manoeuvre his king back to c4 and then go to b4 and a4 after having put the queen on d1 but it is the same draw as we have already seen but from another perspective. This is an important fortress to remember as is the cooperation displayed by the bishop and knight. Notice that the knight is placed on a square of the same colour as the bishop and thus controls squares of an opposite colour. This is a useful construction to remember in the other phases of the games as well. The reason for its importance is the harmony displayed by the three forces, especially the one between the bishop and the knight. The ending in itself is very unusual. I have had it only once in my life when I was a junior and I managed to win with the queen. 283 Lolli 1763 White to move Queen against two bishops is a draw if the attacker doesn’t manage to disrupt the harmony in Black’s set-up or penetrate with his king. 1 ♕d7+ ♔f8 1...♗f7? is a serious mistake since White immediately disrupts Black’s forces with 2 ♔f5. After 2...♗c3 the shortest variation to win is 3 ♕a7! ♗b2 4 ♕b6 ♗a3 5 ♕b7! ♗c5 6 ♕b2+ ♔f8 7 ♕b8+ ♗e8 8 ♔e6 and Black’s c5-bishop is lost in a few moves. The other king moves 1...♔g8 and 1...♔h8 hold the draw, but not 1...♔h6? since the harmony in Black’s position becomes disrupted. This is clearly apparent after 2 ♕c7 with a forced mate after 22 moves. If you are interested in studying how to win in more detail you can consult Nalimov’s Tablebase. 2 ♕e6 ♔g7 3 ♔f4 3...♗h7! This is the only move to keep the balance. 4 ♕d7+ ♔g8 4...♔g6 is also draw but it’s more complicated. It’s best to defend according to the kiss (“Keep it simple stupid!”) principle. 5 ♕g4+ 5 ♔g4 is answered by the only move 5...♗g6. This set-up is very easy to remember for the defender. It’s impossible for White to penetrate with the king or disrupt the forces with his queen. 5...♔f7 If White makes a neutral move Black plays the bishop to g6 and the well-known set-up is reached. 284 Philidor 1777 White to move To win here White must reach the same position with Black to move and create a zugzwang position. This is how to do it: 1 ♕e5+ ♔a7 2 ♕a1+ ♔b8 3 ♕a5 With the help of a triangular manoeuvre (The movement from a5 to e5 to a1 back to a5 is in the shape of a triangle.) Black is forced to move the rook away from the king and then White wins the rook by force with the help of checks. For example: 3...♖b1 4 ♕e5+ ♔a7 5 ♕d4+ ♔a8 6 ♕h8+ 6...♔a7 The only way to avert loss of the rook is by 6...♖b8 but then follows 7 ♕a1 mate. 7 ♕h7+ and White wins the rook next move followed by a quick mate. 285 Grigoriev 1917 Black to move In this position by Grigoriev White holds the draw by oscillating between h3 and e3. Black can never penetrate the third rank with his king and it’s not possible to create a zugzwang situation unless the defender helps you. 1...♔f4+ 2 ♔h2 ♕g4 3 ♖e3 ♔g5 4 ♖h3 ♔h5 5 ♖e3 ♕f4+ 6 ♔g2 ♔g4 7 ♖b3 ♕d2 8 ♖e3 ♕d5+ 8...♕d4 9 ♖h3 is a draw. Be careful not to make the same mistake as Timman did against Nunn in Wijk aan Zee 1982 where White played 9 ♔h2?? and had to resign after 9...h3. White is in zugzwang and there is no salvation after either 10 ♔g1 ♕d1+ 11 ♔h2 ♕f1 or 10 ♖g3+ ♔f4. 9 ♔h2 ♕c6 9...♕c4, threatening a zugzwang after 10...♕f1, is obviously prevented by 10 ♔g2. 10 ♖h3 ♕h6 11 ♔g2 The position is a draw and it’s an important one to remember. 11 ♖e3 allows 11...h3 but that doesn’t matter after 12 ♖g3+ ♔f4 13 ♖xh3 with an easy draw. 286 Averbakh – Bondarevsky Moscow 1948 Black to move Chéron discovered that it was possible to disrupt the harmony in White’s position with the ingenious 60...♕g6+!. 60...♔e6? was actually played by Bondarevsky and leads to the famous study in 1917 by Grigoriev, who discovered that this kind of position is a draw when the defender controls the third rank with the rook. The game continued 61 ♖h3 ♕e4+ 62 ♔h2 ♔f6 63 ♖e3 ♕d5 64 ♖g3 h5 65 ♖e3 ♔g5 66 ♖g3+ ♔f4 67 ♖e3 h4 68 ♖h3 ♕b7 69 ♖e3 (69 ♖xh4+? ♔f3) 69...♔g4 70 ♖h3 ♕b1 71 ♔g2 ♕b7+ 72 ♔h2 ♕c7+ 73 ♔g2 ♕c2 74 ♖e3 and the players agreed to a draw. 61 ♔h2 61 ♔f1 ♕f5 Black prevents the rook from reaching the third rank. 62 ♖h2 ♕f3 63 ♔g1 ♕d1+ 64 ♔g2 ♕g4+ 65 ♔f1 h5 66 ♖h1 (66 ♖g2 ♕d1 mate.) 66...h4 67 ♖g1 ♕d1+ 68 ♔g2 h3+ 69 ♔h2 ♕f3 and Black wins the f2-pawn. 61...♕f5 62 ♔g3 62 ♔g2 ♕g5+ 63 ♔h3 ♕g1. 62...♕e5+ 63 ♔f3 63 ♔h3 ♕g5 64 ♖g4 ♕f5 65 ♔g3 h5 66 ♖h4 ♕g5+ 67 ♔h3 ♕g1 68 ♖xh5+ ♔e4 White has lost the harmony in the position and is completely lost. 63...♕g5 64 ♖h3 64 ♖g4 ♕h5 65 ♔g3 ♕h1. 64...♔d4 65 ♖g3 65 ♖h2 h5 66 ♖g2 ♕h4. 65...♕d5+ Black exploits the white king which isn’t placed behind the rook where it belongs. 66 ♔e2 ♕h1 Black has everything prepared for the h-pawn to queen. 287 Lolli 1763 White to move A relatively common scenario is when the defender has queened a rook pawn and the attacker has the more active position and is to move. If the king and queen are close enough they can construct a mate. 1 ♕d1+ ♔b2 1...♔a2 loses more quickly after 2 ♕a4+ ♔b2 3 ♕b4+ ♔a2 4 ♔c2. 2 ♕c2+ ♔a3 3 ♕c5+! This is the idea to remember. Black loses his queen or succumbs to mate. 3...♔a2 4 ♕c4+ ♔a3 5 ♕a6+ ♔b2 6 ♕b5+ The queen approaches closer and closer in a staircase-like manoeuvre. 6...♔a3 6...♔c1 7 ♕c4+ ♔b2 8 ♕b4+. 7 ♕a5+ ♔b2 8 ♕b4+ ♔a2 8...♔c1 9 ♕d2+ ♔b1 10 ♕c2 mate. 9 ♔c2 ♕c3+ The last trick should always be played out! 10 ♔xc3 and White mates next move. 288 Carlsen – Aronian Elista 2007 White to move In the old days when adjournments were the order of the day this position might very well have been agreed drawn without further play. When Black’s king controls the blockading squares of the pawn it’s a theoretical draw. However, with today’s time control you should play on in this ending. It’s not so easy to hold the draw. Just see what happened to Aronian when he had to defend this drawn ending against Carlsen. 69 ♔h3 ♕b1 It looks simpler to centralise the queen with 69...♕d4 but with little time on the clock and the onset of fatigue it’s easy to get stressed out over the various possibilities of a queen exchange. 70 ♕f6+ ♔h7 71 ♕f3 Note that Carlsen is in no hurry since it is more difficult to defend than to attack. 71...♔g7 72 ♔h4 White avoids moving the g-pawn for as long as possible so as to be sure to win a possible pawn ending if the opportunity should arise after an exchange of queens. Aronian’s plan is simply to stomp with his king. 72...♔h7 The continuation 72...♕h7+ 73 ♔g3 ♕g6+ 74 ♕g4 ♔h8! was playable but it only improves White’s position after 75 ♔h4. However, the position is still a draw. 73 ♕f7+ ♔h6 74 ♕h5+ ♔g7 75 ♕g4+ ♔h7 76 ♕h5+ ♔g7 77 ♕e5+ ♔h7 78 ♕d5 ♔g7 79 ♕d4+ ♔h7 80 ♕d7+ ♔h8 80...♔h6 81 ♕c6+ ♔g7 82 ♔g5 threatens ♕f6+ and a subsequent ♕g6+ but Black easily avoids this by placing the black queen on an appropriate square. 81 g4 Carlsen advances his pawn at the right moment after having shown much patience. Now the black king is in the corner which rather resticts its choices. 81...♕h1+ 82 ♔g5 82...♕c1+?? This fatal check is the losing move. Centralisation with 82...♕e4 was still the key to holding the draw. An amusing variation was 82...♕a8 83 ♕f5 ♕g8+ 84 ♔h6 ♕g7+ 85 ♔h5 ♕f7+ 86 ♕xf7 benefiting from having the king in the corner. 83 ♔g6 ♕b1+ 84 ♕f5 and Aronian resigned since Black cannot prevent the exchange of queens or mate. It’s useful for the defender to remember that it is very dangerous to allow the white king to travel too far down the board since then precise defensive play is required from Black. Don’t forget to practice this ending against a computer. It’s not at all easy with limited time. 289 Ribli – Spassky Montpellier 1985 White to move One of the many revolutionary discoveries of Tablebase was that it analysed two extra pawns on the rook and knight file as a theoretical draw. Here White played 83 h6 with the threat of mate. Ribli’s suggestion of 83 ♕e4, planning g3-g4, is met simplest by 83...♕c3 with a draw according to Tablebase. 83...♕b2 83...♕xh6+? leads to a lost pawn ending after 84 ♕h4 ♔g7 (or 84...♕h7 85 ♔g4! ♕xh4+ 86 ♔xh4 ♔g8 87 ♔g5 ♔g7 88 g4) 85 ♕xh6+ ♔xh6 86 ♔h4 ♔g6 87 ♔g4. 84 ♔h4 84 ♕g7+ ♕xg7 85 hxg7+ ♔xg7 is a draw, since White cannot gain the opposition. However, it would have been a different story with the pawn on g2, but in that case Black would play differently. 84...♕h2+ 85 ♔g5? The position is a draw after the better 85 ♕h3 but why make it so easy for the defender? According to Tablebase Black draws with 85...♕f2! (85...♕d2? loses after 86 ♕e6 ♕d8+ [86...♔h7 87 ♕f7+ ♔xh6 88 ♕h5+ ♔g7 89 ♕g5+ and White wins the pawn ending.] 87 ♔h5 ♕d1+ 88.g4 ♕h1+ 89 ♔g5 [Of course not 89 ♔g6? ♕xh6+ 90 ♔xh6 with a stalemate.] 89...♕c1+ 90 ♔f5! ♕c2+ 91 ♕e4 ♕f2+ 92 ♔g5! ♕d2+ 93 ♕f4 ♕d5+ 94 ♕f5 ♕g8+ 95 ♔h4! and White wins.) The point is that after 86 ♕e6 ♔h7 White cannot force the exchange of queens. 85...♕xh6+ The players agreed to a draw due to stalemate. The best and only way to learn how to defend this endgame is to play it against the computer. 290 Ciocaltea – Unzicker Moscow 1956M Black to move Black should play according to a three-step-plan: 1) Centralisation of the queen. 2) Advance of the king into enemy territory. 3) Advance of the f-pawn to decide the game. The game continued: 52...♕c1+ 53 ♔h2 ♕f4+ 54 g3 Pachman and Averbakh believe Black’s winning chances increase after this move, but White still keeps the draw with correct play. As a matter of a fact, with the pawns placed on g3 and h4 White holds the draw without any problems, but the f3-square should be looked after as Black can exploit it if White isn’t careful. According to Pachman the golden rule in queen endings is the security of the king and that is the reason White should prefer 54 ♔h1. Averbakh also prefers this move as the safer option. The drawback is that Black has an alternative plan to the advance of the king and that is to push the g-pawn to g3, creating mating threats on the last rank. 54 ♔g1 is a recommendation by Giddins but the king is slightly more exposed here for timely central checks along the classical diagonal (a7-g1), so the recommendation by Pachman and Averbakh is more precise. 54...♕f3 The queen stands in an excellent attacking position and is defending the f7-pawn as well. 55 h4 h5 According to Dvoretsky the position is a draw if White doesn’t allow the black king to penetrate the white position. 56 ♕d2 ♔g7 57 ♕d4+ ♕f6 58 ♕d2 58...♕e5 Centralisation of the queen is a very important principle in queen endings as it makes it more difficult for the opponent to give annoying checks. Step 1 has been achieved and Black is preparing for the next step: the advance of the king. 59 ♔g2 ♔f6! Black’s goal is to penetrate the white position and White has to prevent it. 60 ♕d8+ ♔f5 61 ♕d7+ ♕e6 62 ♕b5+ ♕e5 62...♔e4? leads nowhere after 63 ♕e2+ ♔d4 64 ♕d2+ ♔c4 65 ♕c2+ ♔b4 66 ♕d2+ ♔a4 67 ♕d4+. 63 ♕d7+ ♕e6 64 ♕b5+ ♔f6 65 ♕g5+ ♔g7 66 ♕d2 So far White has defended well, therefore Black has to try something else. 66...♕c4! A good spot for the queen which once again prepares the advance of the king. 67 ♔h2?! White plays unnecessarily passively when other, simpler, options were available such as 67 ♕b2+, 67 ♕d6 or 66 ♕d8, effectively stopping Black from activating his king. This mistake was probably connected with an underestimation of the value of an active king for the attacker in this type of queen ending. 67...♔f6! 68 ♕d8+ ♔e6 69 ♕e8+ ♔f5 The weak link in Black’s pawn chain, the f7-pawn, is defended by the centralised queen. This is a necessary condition for Black to be able to advance his king. 70 ♕d7+ ♔e4 71 ♕e7+ ♔d3 The key idea for Black is to place his king near the white king. This will make it possible for Black to place his queen on the same rank as the white king, thus nullifying the effects of checks. 72 ♕a3+ ♔c2 73 ♕e7? 73 ♕d6? would have prevented the black king from passing the d-file but it would have still lost due to 73 ...♕e2+ followed by the exchange of queens. For example 74 ♔h3 ♕e6+ 75 ♕xe6 fxe6 with a won pawn ending. 73...♔d1 74 ♕d8+ ♔e2 75 ♕e7+ ♕e6 76 ♕b7 ♔f2! Now Black’s king is close enough to the opponent’s king. 77 ♕g2+ 77 ♕b2+ ♕e2 is the typical idea Black is striving for. Black plays for mate or the exchange of queens. 77...♔e1 78 ♕g1+ ♔e2 79 ♕g2+ ♔d3! 80 ♕f3+ ♔d2 81 ♕f4+ ♔e2 82 ♕c7 It’s time for step 3. 82...f5 The defensive possibilities are worse when the kings are close to each other but White can still hold the draw with precise play. 83 ♕c2+ ♔f3 84 ♕g2+ ♔e3 85 ♕b2 If 85 ♕g1+ ♔d2 (Here Giddins recommends 85...♔e2 86 ♕g2+ ♔d3 87 ♕f3+ ♔d2 88 ♕f4+ ♔e2 89 ♕c7 ♕e4! but after 90 ♕c1! the position is still a draw.) 86 ♕d4+ ♔e2 87 ♕b2+ ♔f3 88 ♕g2+ (88 ♕c3+? is answered by 88...♕e3! Don’t forget this important idea! [88...♔f2? is wrong due to 89 ♕d2+ ♕e2 90 ♕f4+ ♔e1+ 91 ♔g1!] 89 ♕c6+ ♕e4 90 ♕c3+ ♔f2 91 ♕d2+ ♔f1! 92 ♕c1+ ♕e1 and Black wins. Note that Black’s queen should be placed on White’s first or third rank to benefit from the crosscheck technique!) 88...♔e3 89 ♕g1+ is a draw. 85...♕c4 86 ♕a3+ ♕d3 87 ♕c5+? 87 ♕a1! is a draw according to Nunn and so is 87 ♕a7+. 87...♔f3 Black is now winning. 88 ♕c6+ 88 ♕g1 f4! The culmination of step 3 destroys White’s position completely. 89 gxf4 (If 89 ♕g2+ ♔e3 90 ♕g1+ ♔d2 91 gxf4 ♕e2+ Black wins the forthcoming pawn ending due to his far more active king.) 89...♕c2+ 90 ♔h1 ♕e4! 91 ♕g5 ♕b1+ 92 ♔h2 ♕c2+ 93 ♔h1 ♕c1+ 94 ♔h2 ♕xf4+ and Black wins according to an analysis by Unzicker. 88...♕e4 89 ♕c3+ ♔f2 90 ♕c5+ ♕e3 91 ♕c2+ ♕e2! Now it works to place the queen on the second rank and apply the cross-check technique, since the check on f4 isn’t available. 92 ♕c6 ♔f1+ 92...f4 93 gxf4 ♔e3+ according to step 3 also wins easily. 93 ♔h3 ♔g1! The culmination of the long king march. 94 ♕c5+ ♕f2 95 ♕e3! f4!! White resigned. The culmination of step 3 was conducted in the most devastating and beautiful manner. 95...♕xe3?? would obviously have led to a stalemate. White’s last hope had vanished. A very instructive endgame showing that the more active the attacker’s king, the more difficult the defence. 291 Alekhine – Reshevsky AVRO tournament, Holland 1938 White to move According to Averbakh the defender has good drawing chances when, as here, the attacker has weakened his king’s position. The second rank is weak and White’s king position would have been safer with the pawn on f2 instead of f3. 44 ♕a2 ♔g8 Black defends the weakest link in his pawn position while sidestepping the long dark-squared diagonal. 45 a4 ♕c6 46 a5 ♕a6 In contrast to rook endings, the defender can temporarily place his queen in a passive offside position in front of the passed pawn. There will normally be room for annoying checks later when White strives to go to b6 with the queen. In rook endings it’s easy to start advancing the king but here, for the time being, it has no effect. 47 g4 White improves his position on the kingside before activating his queen. 47 ♕d5 allows 47...♕e2+ 48 ♔h3 h5 with a positional draw since it’s not possible for White to improve his position. For example if White plays 49 ♕e4 Black replies 49...♕f1+ 50 ♔h4 ♕b5 and it’s not possible to retain the passer. An amusing trap is 51 ♕a8+ ♔g7 52 a6? ♕e2 53 ♔h3? ♕f1+ 54 ♔h4 ♕g2 and Black wins. One of the main differences between queen and rook endings is that in the former the attacker must be more careful when playing for a win. 47...g5 48 ♔f2 ♕d6 49 ♔f1 ♕a6+ 50 ♔g2 ♔g7 51 ♕b2+ ♔g8 52 ♕b8+ ♔g7 53 ♕e5+ ♔g8 54 ♔f2 ♕a7+ 55 ♔e2 ♕a6+ 56 ♔d2 ♕c4! Black cuts off the white king so it will not be able to enter the cfile. If White manages to establish cooperation between his king, queen and a-pawn there is a significantly higher risk that Black will lose. 57 ♕f5? This is a mistake after which the position is a draw by force. A better try was 57 ♕c3 or 57 h3. 57...♕d4+ 58 ♔e2 ♕b2+ 59 ♔d3 ♕b3+ 60 ♔e2 ♕b2+ Drawn. The main rule for queen endings with a distant passed pawn is that your king’s position is relatively safe or that you can establish cooperation between your pieces. If not, then this type of endgame should normally end peacefully. 292 Alekhine – Euwe Nottingham 1936 White to move In the tournament book Alekhine comments that the plan is to advance the pawns to a5 and f4 at the right moment. However, there is only one way to win in the long run and that is by a timely activation of the king. 56 ♕b4 ♕h1 56...♕a2+ actually helps White to improve his positon with 57 ♔e3 and the further checks 57...♕e6+ 58 ♔d3 ♕a6+ lead nowhere after 59 ♕c4 since 59...♕d6+ is not possible after 60 ♕d4+. 57 ♕e1 ♕h2+ 58 ♔e3 ♔h7 59 a5 ♕a2 60 ♕d2 ♕a1 61 ♔e2 ♔h6 61...♕e5+ 62 ♔f2 ♕c5+ 63 ♔g2 ♕c4 64 f4 leads to the same kind of play as in the game. 62 f4! White’s play is very instructive. White wants to play a pawn to g5 and then drive the black king to the eighth rank. Then White advances his king towards the passed pawn. 62...gxf4 63 gxf4 ♕a4 64 ♔f2 White is in no hurry to play 64 g5+ but in fact it was possible since 64...♔h5 is answered by 65 ♔f3! ♕a3+ (65...♕c6+ 66 ♔f2) 66 ♕e3 ♕xa5 67 ♕e2!! (67 ♔g3?? leads to stalemate after either 63...♕xg5+! or 63...♕e1+!.) 67...♕a3+ 68 ♔g2+ ♔h4 69 ♕f3 ♕a2+ 70 ♕f2+ ♕xf2+ 71 ♔xf2 and White wins the pawn ending. 64...♔h7 65 g5! ♕a3 Black wants to cut off the white king on the third rank. 66 ♕d7+ ♔h8 67 ♕c8+ ♔h7 68 ♕c7+ An excellent square for the queen, where it performs three tasks: Black’s king is cut off and both pawns are protected. 68...♔h8 69 ♔e2 At last White’s king approaches the a-pawn. 69...♕a2+ A waiting policy doesn’t work after 69...♔g8 70 ♔d2 ♔f8 (or 70...♔h8 71 ♕c3+) 71 ♕d8+ ♔g7 72 ♕f6+ ♔h7 73 a6 since Black is nevertheless forced to check the white king. However, this will be of no avail since White will find protection on the eighth rank after all the checks are over. 70 ♔e3 ♕b3+ 71 ♔d4 ♕b4+ 72 ♔d5 ♕b5+ 72...♔g8 is answered by 73 a6! when the pawn cannot be taken after 73...♕b5+ 74 ♔d4 ♕xa6 due to 75 ♕c4+. The arising pawn ending is so strong for White that it would even be won without the f-pawn. 73 ♔d4 73 ♔d6 was more precise, for example 73...♕d3+ 74 ♔e7, but Alekhine wanted to win in the safest manner rather than in the quickest way. 73...♕a6 73...♕b4+ 74 ♔e5 ♕b5+ 75 ♔f6 ♕f5+ 76 ♔e7 ♕f8+! is a typical stalemate idea for Black and may explain why he persists in keeping his king in the corner. However, this isn’t a draw since White can play 77 ♔d7 with a win, but it proves that it’s not easy for the king to catch the g6-pawn. 74 ♕b6 ♕c8 75 ♕d6! Perhaps this is an over-refinement since 75 a6 was perfectly playable. Nevertheless the centralisation of the queen is good technique and it’s useful to remember it. Obviously not 75 ♕xg6?? when Black can check the white king any way he wants to, without fearing the capture of the queen. 75...♕c2 76 a6 ♕d2+ 77 ♔e5 ♕c3+ 78 ♔e6 ♕c8+ 79 ♔e7 Everything goes according to plan and now Euwe must defend against the check on the eighth rank. 79...♔h7 80 ♕d7 ♕c3 81 ♔e6+ and Black resigned since there will be an exchange of queens on the f6-square. 293 Sämisch – Maróczy Carlsbad 1929 Black to move According to Fine and Dvoretsky this ending is a draw. Speelman thinks White has good winning chances. Whatever the great endgame experts say, everyone should be careful in this type of ending because of the presence of queens. It’s actually possible to lose this ending even in high level chess as once happened to Larsen in a game he played against Keres with the white pieces, as can be seen in position 298. 46 ♕c7+ ♔g8 47 f3 ♕d3 48 e4 ♕d2 49 ♔h2 ♕b2 50 ♕g3 ♔f7 51 f4 This pawn advance doesn’t necessarily prepare e4-e5 but it does prepare f4-f5 since Black had weakened his position with ...h5 earlier in the game. 51...♕e2 52 ♕b3+ 52 e5?! doesn’t lead to anything substantial after 52...♕e4. Soon White is forced to capture on f6 thus making Black’s defensive task easier. 52...♔f8? Maróczy, who was a specialist in handling queen endings, makes an uncharacteristic mistake. 52...♔e7 should have been played. 53 ♕b8+? Surprisingly White overlooks the strong and natural move 53 ♕e6!, preparing e4-e5 while threatening the f6-pawn with check. 53...♔f7 54 ♕c7+ ♔e6 The only move, otherwise White pushes the e-pawn with great effect. 55 f5+ gxf5 56 ♕c8+ ♔e7 57 ♕xf5 Although Black’s pawn position is scattered it’s strong due to the fact that it’s impossible for White to create a passed pawn. 57...♔f7 58 ♔h3 ♕d3+ 59 ♕f3 ♕d7+ 60 ♔h2 ♕b5!? 60...♔g6 looks more to the point since White now gets his passed pawn. However, the position is still a fairly easy draw. 61 g4 hxg4 62 ♕xg4 Although White’s passed pawn is a rook pawn, Black has to be careful not to allow the white king to penetrate his position. 62...♕b2+ 63 ♔g3 ♕c3+ 64 ♕f3 ♕c7+ 65 ♔g4 ♕d7+ 66 ♔h5 ♕e8! A good move which eyes the slightly exposed white king. 67 ♕f5 ♕h8+ 68 ♔g4 ♕g8+ 69 ♔f4 69...♕g1? A crucial mistake allowing White to advance his king deeper into black territory. 69...♕g2 would not have allowed that due to 70 ♕d7+ ♔f8 71 ♔f5?? ♕h3+ and Black wins. An even simpler draw was 69...♕b8+ with the idea of 70 e5 ♕b4+ 71 ♕e4 fxe5+ 72 ♔xe5 ♕xe4+ 73 ♔xe4 ♔g6. 70 ♕d7+ ♔f8 70...♔g6 is out of the question after 71 ♕g4+ leading to a winning pawn endgame. 71 ♕d5? 71 ♔f5 ♕f1+ 72 ♔g6 was the most logical move to play for a win. After 72...♕g1+ 73 ♔xf6 ♕b6+ 74 ♕e6 ♕f2+ 75 ♕f5 Black has no worthwhile continuation except to resign. 71...♕g6 Black prevents the further advance of the enemy king. 72 h5 ♕h6+ 73 ♔g4 ♔e7 The simplest draw was 73...♕g7+ 74 ♔f5 ♕h7+ 75 ♔xf6 ♕h6+ and the h5-pawn falls with check. 74 ♕c5+ ♔d7 75 ♕f5+ ♔e7 76 ♕f4 ♕g7+ 77 ♔h4 ♕g2 78 h6 ♔f7 79 ♕c7+ ♔g8 80 ♕d8+ ♔h7 81 ♕e7+ ♔xh6 82 ♕xf6+ ♔h7 83 ♕f7+ ♔h8 84 ♕h5+ ♔g8! 84...♔g7? 85 ♕g5+ ♕xg5+ 86 ♔xg5 and White wins thanks to the opposition. 85 ♕f5 The position is an elementary draw but for some reason Maróczy lost at this moment. Presumably he lost on time. Note that 85 ♕g4+ ♕xg4+ 86 ♔xg4 ♔f8! would have led to another important position presented in the section on pawn endgames. 294 Reshevsky – Fine Nottingham 1936 Black to move In the tournament book Alekhine thinks that Black has an easy defence if he walks with the king to the centre and thereby halts the future passed pawn on the d-file. White’s exposed king will also help Black and give him counterchances. Fine doesn’t agree and thinks White has good winning chances. Normally the winning chances are greater the more pawns you have. It should be easier to play for a win with five versus four compared with four versus three. 38...♕a1 39 ♕e5+ ♔g8 40 ♔g2 White prevents the black queen from occupying the other corner. In such a situation the h2 pawn would have had to move, thus creating a glaring weakness on g3. 40...♕e1 A strong position for the black queen in the heart of White’s position. 41 ♕b8+ ♔g7 42 ♕b2 ♔g8 43 ♕c2 White’s clever queen manoeuvre plans e2-e4 followed by d4-d5 and Black has to react to this by trying to blockade the pawn on d5. 43...♕b4 44 ♕d3 It’s sensible to place the queen behind the future passer. 44...♔f8 45 ♔h3 ♕a5 46 e4 ♕h5+ 47 ♔g2 ♕a5 48 d5 ♔e7 49 ♕d4! Placing the queen on one of the long diagonals is an important device to master in queen endings. We’ll touch on this idea again in the queen ending between Kramnik and Adams, position 195. 49...exd5 50 exd5 ♔d6 Fine plays aggressively, according to Alekhine’s plan, but it’s not as easy to draw as Alekhine thought. 51 ♕f6+ 51...♔xd5 The clever 51...♔c5!! was undoubtedly not easy to find over the board. The main variation goes 52 ♕e7+ ♔b5 53 d6 ♕d2+ 54 ♔h3 ♕h6+ 55 ♕h4 (55 ♔g4?? f5 mate) 55...♕d2 56 ♕xh7 ♕d5! 57 ♔h4 ♕xf3 58 ♔g5 ♔c6 and Black achieves a draw. This variation proves that Alekhine was right with his ingenious comment that Black should play aggressively with the king, but the variations themselves are not easy to master. 52 ♕xf7+ ♔d4 52...♔c6 was more precise since 53 ♕xh7 is met by 53...♕a2+ 54 ♔h3 ♕e6+ 55 g4 ♕f6! 56 ♕h4 ♕xf3+ 57 ♕g3 ♕f6 with some drawing chances. 53 ♕xh7? The queen becomes displaced here and Black’s activity creates serious counterplay. Correct was 53 ♕d7+! ♔e3 54 ♕e7+ ♔d3 (54...♔d4? 55 ♕e4+ forces the exchange of queens with an easy win.) 55 ♕e4+ ♔d2 56 f4 with good winning chances since the black king is cut off while White improves his position with every move. 53 ...♕a2+ 54 ♔h3 ♕e6+ 55 g4 ♔e3! 56 ♕b7 ♔f2! 57 ♕b8 ♔xf3 58 ♕f8+ ♔e2 59 ♕f4? 59 ♔h4! was more ambitious. 59...♕d5! Black’s queen is too strong on this centralised square and, as there is no way for White to improve his position, the players agreed to a draw. 295 Kramnik – Adams Wijk aan Zee 2000 White to move In this exciting position Kramnik has two advanced passed pawns and Black one. The most straightforward win is 45 d6! b3. After 45...♕xd6 46 ♕e4+ we have the following branches: a) 46...g6 47 ♕b7 The value of this position is that it shows the importance of the long white diagonal. With the help of this diagonal White can protect not only the king from annoying checks but also support the a-pawn. b) 46...f5 47 ♕b7. c) 46...♕g6 47 ♕xb4 ♕c6+ 48 ♔h2 (48 f3? leads to a similar situation that occurred in position 291 between Alekhine and Reshevsky, who managed to draw due to the weakened king’s position.) 48...♕f3 49 ♕b6 and White wins easily. 46 d7 ♕c6+ Black understands the importance of the long diagonal but succumbs to the passed pawns. 47 ♔h2 ♕f3 48 ♕e1 ♕d3 49 a6 and Black resigned realising that 49...♕xd7 50 ♕b1+ was a hopeless task with White’s secured king. 296 Rubinstein – Capablanca Saint Petersburg 1914 Black to move Black is a pawn down and the c5-pawn is under attack. White’s queen is more active than Black’s. Black’s only chance of staying alive is to exploit the pawn majority on the queenside to create a passed pawn. In queen endings one of the golden rules is that it is more important to have a passed pawn than a material advantage of one or even several pawns. 27...b4! An excellent move which exploits the fact that it is easier for Black to create a passed pawn than it is for White. 28 ♕xc5 Other options were: a) 28 cxb4 which is met by 28...♕xb4! 29 ♕xa6 c4 when Black’s passed pawn is far more dangerous than White’s. 30 ♕a8+ ♔h7 31 ♕e4+ ♔g8 32 ♕e8+ ♔h7 33 ♕xf7 c3 forces White to go for perpetual checks because of the deadly passed pawn. b) 28 c4!? was perhaps a better winning attempt. Nevertheless Black has good drawing chances after 28...♕c8 (Tarrasch’s recommendation 28...♕a7, preparing ...a5, ...a4 and ...b3, can be met by 29 ♕d8+ ♔h7 30 ♕a5 but then Black has the strong 30...♕e7 31 ♕xa6 ♕e4 with a positional draw. It’s not possible for White to move his queen without losing one of his queenside pawns, giving Black a passer, and if White plays 32 h5 Black can play a waiting game with 32...♕b1+ 33 ♔h2 ♕e4.) 29 ♕b6 ♕f5!. Activity is important not only in rook endings but also in those with queens. Black will not passively defend his pawns but strive at all cost to get a passed pawn on the queenside. 30 ♕xa6 ♔h7! Black is two pawns down but still has time to secure his king position and evacuate the eighth rank to avoid giving White a possible check with a gain of tempo. 31 ♕a7 f6! Otherwise White would play e3-e4 next move, forcing Black to give up either the c5-pawn or the f7-pawn. 32 ♕e7 ♕b1+ 33 ♔h2 ♕xa2 34 ♕e4+ ♔h8 35 ♕e8+ ♔h7 36 ♕h5+ ♔g8 37 ♕xc5 b3 Black’s passed pawn is too dangerous and sooner or later White will be forced to go for a perpetual. 28...bxc3 29 ♕xc3 ♕b1+ 30 ♔h2 ♕xa2 Black’s passed a-pawn is much more dangerous than White’s four-to-two-majority on the kingside so White is forced to look for perpetual checks. 31 ♕c8+ ♔h7 32 ♕f5+ g6 33 ♕f6 a5 34 g4 a4 35 h5 35...gxh5 35...a3 looks like mate after 36 h6 but Black has the saving move 36...♕b2! (Not 36...♔xh6? 37 ♕h8+ ♔g5 38 ♔g3 and Black must give up his queen to avoid mate next move.). 36 ♕f5+ If White plays too aggressively with 36 gxh5 Black replies 36...♕d5!. Now 37 h6? isn’t possible because of 37...♕h5+ 38 ♔g3 ♕g6+ and Black wins. 36...♔g7 37 ♕g5+ ♔h7 38 ♕xh5+ ♔g7 Here the players agreed to a draw. In the main variation we saw that Black obtained a dangerous passed pawn on the a-file, while in the sidelines Black got passers on the bor c-file. It doesn’t matter on which file the passer is created, more important is that it’s an advanced passed pawn which can neutralise a material disadvantage and offer a balanced position – and for this a passed pawn is necessary in order to force the opponent to go for perpetual checks. If it’s not possible to create a passed pawn, then another way to achieve good drawing chances is to exchange all the pawns on the queenside and try to hold with, say, three pawns against four on the kingside. The present work classifies the next four queen and rook endgames as “NQE” (“Not Quite an Endgame”), pronounced “nuckie” to use the term introduced by Flear, and I thereby go against the definition of Speelman who says that this should be regarded a middlegame. After all, queen and rook has the value 9 + 5 = 14 and not 13 which is Speelman’s clever definition of an endgame. However, this endgame is obviously a borderline case. 297 Flear – Lejeune France 1995 Black to move If one side has an extra pawn but a more exposed king the situation is often unpredictable. In practice this means it’s often fine to be a pawn down if the opponent has the worse king’s position. 36...♖c1 37 ♕d8+ 37 ♕xa7 ♕g5 looks dangerous but according to the computer the position is balanced after 38 ♔e2 and White’s queen is unable to defend the king. 37...♔h7 38 ♕d7 ♕c5+ 38...♕e5? 39 ♕f5+ ♕xf5 40 exf5 is an advantageous rook ending thanks to the pawn majority and the active king. 39 ♕d4?! One can understand White’s desire to exchange queens. Better was 39 ♔g3 ♕e5+ 40 ♔h3 ♕h5+ 41 ♔g2 ♕g5+ 42 ♔h3 ♕h5+ with a draw by perpetual check. 39...♕g5 39...♕h5! was slightly more testing. 40 ♔e2 ♖h1?! The quiet move 40...a5! was better. White’s exposed king will not run away. Black is active and this prophylactic move targets the white pawn majority. The dream is to play the pawn to a4. 41 ♔d3 ♕h5 42 ♕xa7 ♕xf3+ 42...♖c1!? points to the fact that the centre of gravity is the king rather than the material. A plausible response is 43 ♕f2 ♕b5+ 44 ♔e3 ♕c5+ 45 ♔e2 ♕c4+ 46 ♔e3 ♕c5+ with a draw. 43 ♕e3 ♕f6 44 ♕d4 ♕f1+ 45 ♔c2 42...♕b1+? The turning point in this middlegame/endgame. 45...♕a6 would have made it more difficult for White to defend his king. 46 ♔b3 (46 ♕d3 ♕a4+) 46...♖a1 denies the white king the relatively safe a2-square and Black can continue his annoying queen checks to keep the balance. 46 ♔b3 ♖f1 46...♖e1 47 ♕d3 ♕xd3+ 48 ♖xd3 ♖xe4 49 a4 wins easily for White. 47 ♕d3 ♕xd3+ 47...♕c1 gives Black’s king problems after 48 e5+ g6 49 e6. Black can reply 49...♖f6 but after 50 ♕c3 ♕xc3+ 51 ♔xc3 ♖xe6 52 a4 the situation is tough for Black. 48 ♖xd3 g5 49 a4 h5 50 a5 g4 51 ♔a2 ♖f6 52 ♖a3 ♖a6 53 b4 h4 54 b5 ♖g6 55 a6 g3 56 hxg3 h3 57 ♖f3 h2 58 ♖xf7+ Black resigned. The result of the game was certainly unpredictable since Black missed several good opportunities. 298 Larsen – Keres San Antonio 1972 White to move Compared with other endgames where the attacker has four pawns versus three on the same side, here the game can be sharp. White’s extra pawn is not so easy to exploit and this particular endgame was even lost by Larsen! That says something about the difficulties in playing with major pieces. You must always take care with the king’s position. 31 ♖d6 g6 Black has set up the wave which is the best defensive structure if White exchanges to a rook ending. As a rule White should avoid the exchange of rooks and in the worst case scenario play a queen ending, but the best is to keep all the major pieces on the board. 32 ♖d7 ♕e6 33 ♖d8+ ♔g7 34 ♕b4 If White finds it necessary to exchange queens then he must at least get a passed pawn, as in the variation after 34 ♕b2+ ♕e5 35 ♕d4. If White concentrates his major pieces on the eighth rank with 34 ♕b8, then Black has a solid defence after the sequence 34...♕e4+ 35 ♔h2 ♖c2 36 ♖g8+ ♔h7 37 ♖h8+ ♔g7 and so on. 34...♕c4 35 ♕b2+ ♕c3 36 ♕b7 ♖f5 37 ♕e7 ♕b2 It’s typical for play with major pieces that there are mutual threats on the king’s position. 38 ♕f8+ ♔h7 Note the strong queen on b2 which both attacks and defends the king. 39 ♕g8+ ♔h6 40 ♖d2 ♕g7 Of course not 40...♕xd2?? 41 ♕h8 mate. 41 ♕a8 ♔h7 41...♕c3 42 ♖d5 ♖xd5 43 ♕xd5 offers better practical winning chances than a rook ending. 42 e4 42...♕c3! This is actually the only move. 42...♖e5? loses after 43 ♖d7 ♖e6 44 ♕b7 ♔g8 (Upon 44...♖f6? the epawn decides with 45 e5 ♖f5 46 e6.) 45 ♕c8+ ♔h7 46 ♖c7! A very nice idea. The immediate threat is ♕xe6. 46...♖f6 (46...♖xe4? 47 ♕b7 loses immediately and 46...♖e5 loses as well after 47 ♕b7 ♔g8 48 ♕b8+ ♔h7 49 ♖xf7.) 47 ♕b7! ♔g8 (47...♔h6 48 e6 ♖f5 49 f4 – not 49 e6? ♕f6!.) and the rook is trapped – 48 e5 ♖e6 49 ♕c8+ ♕f8 50 ♕d7 ♕g7 (50...♖e8 defends the last rank but after 51 e6! Black’s weaknesses on the seventh and eighth ranks are clearly visible.) 51 ♕xe6! fxe6 52 ♖xg7+ ♔xg7 53 ♔f3 and White wins the pawn ending. 43 exf5 ♕xd2 44 ♕b7 ♔g8 45 f6 45 fxg6 fxg6 46 ♕e4 obviously gives White fewer winning chances. 45...♕d6 46 ♕c8+ ♔h7 47 ♕c3 ♕d5+ 48 f3 ♕a2+ 49 ♔h3 ♕b1 50 ♔g2 ♕a2+ 51 ♔f1 ♕a6+ 52 ♔e1 ♕e6+ 53 ♔f2 ♕a2+ 54 ♔g2 ♕b1+ 55 ♔g2 ♕a2+ 56 ♔h3 ♕b1 57 g4 ♕h1+ 58 ♔g3 ♕g1+ 59 ♔f4? On the surface it’s hard to believe that this activation of the king is a mistake but it is, and a very instructive one to remember. It’s necessary to accept the draw after 59 ♔h3. 59...♕h2+ 60 ♔g5? 60 ♔e3 ♕xh4 61 gxh5 gxh5 62 ♕d3+ ♔h6 63 ♕d8 was a better chance to draw but admittedly the f6-pawn is a weakness and the distant passed pawn dangerous, so a draw is still not easy to achieve. 60...♕g3! After this strong move White is completely lost. It’s incredible but chess is full of such miracles and we must learn them one by one and be prepared for them because a knowledge of the exceptions separates the best from the rest. 61 ♕e3 hxg4 62 ♕f4 ♕xf3 63 ♕xg4 ♕e3+ 64 ♕f4 ♕e2 65 ♕g3 ♕b5+ 66 ♔f4 ♕f5+ 67 ♔e3 ♕xf6 68 ♕g5 ♕f1 Larsen plays the best defensive moves but it’s to no avail. 68...♕xg5+? 69 hxg5 is if course a stone cold draw. 69 ♕g4 ♕e1+ 70 ♔d3 ♕e6 71 ♕f4 ♔g7 72 ♕d4+ f6 73 ♕b4 ♕f5+ 74 ♔e2 ♔h6 White’s king is cut off and Black will soon pick up the h4-pawn. The rest needs no commentary: 75 ♔e1 ♔h5 76 ♕c4 ♕g4 77 ♕c5+ ♔xh4 78 ♕e7 ♕f5 79 ♕b4+ ♔h5 80 ♕c4 g5 81 ♕f7+ ♔h4 82 ♕f8 ♔g3 83 ♕a3+ ♕f3 84 ♕d6+ ♔g2 85 ♕d2+ ♔h3 86 ♕d7+ f5 87 ♕g7 g4 88 ♕h8+ ♔g3 89 ♕e5+ f4 90 ♕b8 ♕e3+ 91 ♔d1 ♔g2 White resigned. Principally the defender should strive for an exchange of queens and if this isn’t possible then the second best is an exchange of rooks. The attacker tries to retain the queen and rook while introducing a third attacker, preferably the e-pawn. The other pawns should protect the king from counterattack. We now turn to two positions where the attacker has a positional advantage. In the first case we are discussing the isolated pawn and in the second how a more active position can be exploited by an unexpected breakthrough in the centre. 299 Giri – Dominguez Perez Novi Sad 2016 White to move Black has just played 28...f5 preventing White from winning a pawn with e3-e4. Such further weakening moves (the second rank becomes vulnerable) are typical of positions with an isolated pawn. When there are only major pieces on the board the isolani actually becomes weaker due to the strong pressure on the d-file. 29 h4! White plans the pawn push to h5 in order to fix the g7-pawn and artificially isolate the f5-pawn. 29...♖d7 30 ♕c3! A good manoeuvre with the queen, exploiting the fact that Black cannot attack it on the c-file with the rook. 30...♔f7 31 h5 The principle of two weaknesses is clearly visible here. Both f5 and g7 are potential future weaknesses, as well as d5 and to a certain extent a7, so in reality Black has four weaknesses. 31...♕f6 32 ♖d4 Notice how White uses the manoeuvring square d4 when mobilising the rook to the fourth rank. The change of pieces on the d4-pivot is typical for manoeuvring when playing against several weaknesses. 32...♔e7 By placing the king in an exposed position Black risks having a fifth weakness consisting of His Royal Highness himself. The safest was to play 32...♔g8 followed by ...♔h7. King safety is the number one thing when playing with queen and rook. 33 a4! White wants to clarify the situation and create a concrete weakness on the queenside. 33...♕e5 34 a5 bxa5 35 ♕xa5 Here it’s clearly visible that Black has four pawn weaknesses. Since White has the more active pieces it’s very difficult for Black to defend them all in practical chess. Black must also consider the safety of his king which certainly must be counted as a fifth weakness. 35...♕e6 36 ♖f4 ♕e5? The best defence was 36...♔f8! with the idea 37 g4? ♔g8 38 gxf5 (38 ♖xf5 d4 with counterplay.) 38...♕f7 39 ♕b5 ♖b7 40 ♕d3 ♖d7 41 ♕d1 ♖d6 42 ♕f3 ♖f6 and White’s pieces are tied to the defence of the kingside pawns. 37 ♕c5+ ♔e6 38 ♕c8 ♔e7 39 ♕c5+ ♔e6 40 ♕c6+ ♔e7 41 ♕g6 ♕f6 Here White must make the important decision whether to play a rook ending or keep queens on the board. 42 ♖xf5! Giri forces a transition to a technically winning rook ending where the g6-pawn provides an outpost on f7. The alternative was to play 42 ♕h7 ♔e6 43 ♖a4 (43 ♕g8+ ♕f7) and continue manoeuvring against Black’s many weaknesses. Black can try to hold on with 43...♕f7. 42...♕xg6 43 hxg6 ♔e6 44 ♖f8 h5 45 ♔g2 d4 46 exd4 ♖xd4 47 f4 ♖d5 48 ♖f7 ♖b5 49 ♖xg7 ♔f6 50 ♖xa7 ♔xg6 51 ♖a6+ ♔g7 52 ♔h3 ♖xb3 53 ♖a5 Black resigned. 300 Rozentalis – Engqvist Taby 2009 White to move I must confess that I had completely overlooked the ingenious and seemingly impossible move: 51 d5!!. Next day somebody told Rozentalis that Kasparov had played something similar where Black could take on d5 with three pieces. 51...exd5 51...♕xd5 52 ♖b8+ leads to mate in three moves and 51...cxd5 is answered by the winning 52 ♖d6! but not 52 ♖b8+? ♖c8 53 ♖xc8+ ♕xc8 54 ♕e7 d4 55 ♕d6 d3 56 c6 d2 57 c7 ♕xc7!. Objectively speaking, 51...♖c8 was the best but to play a position where White has a protected pawn on d6 is of course futile in the long run. Black can hardly move and has three weaknesses apart from his passive major pieces: the c6-pawn must be taken care of, the d6-pawn kept under surveillance and especially the f7-pawn must be protected if Black wants to manoeuvre his king to h6. 52 ♖b8+! 52 e6 ♕e8 53 exf7+ ♖xf7 54 ♕xg6+ ♔h8 55 ♕xh5+ ♔g7 56 ♔h2!. 52...♖c8 53 e6 Another strong break... 53...fxe6 54 ♕xg6+ ♔h8 54...♔f8 55 ♖b3 ♔e7 56 ♖f3 (56 ♖b7 ♖c7 57 ♕g7+ ♔d8 58 ♕f8+ ♕e8 59 ♖b8+ is mate in twelve.) 56...♔d8 57 ♖f7. 55 ♕xh5+ ♔g7 56 ♖b4 ♖f8 If 56...d4 57 ♕e5+. 57 ♖g4+ ♔f6 58 ♖f4+ ♔e7 59 ♕h7+ ♔e8 Black resigned. What is fascinating with the theme “impossible breakthrough” is that it’s very unusual. But it has happened twice to me and against the same player! As a matter of fact it already happened in a game in Stockholm the last day of 2006. Rozentalis played what for me was a shocking move, namely 43...b4!?. If I play 44 axb4? Black plays 44…a3 45 bxa3 ♖xc3+ followed by 46…♖xa3. In the game I played 44 cxb4 and after 44… ♔f8 45 ♘d3 ♖d8 the game was in a state of dynamic equilibrium. Nevertheless I managed to lose after 63 moves apparently because I was overly impressed by the impossible move 43…b4!? as well as his faultless technique using the bishop pair. Another thing concerning pawn breakthroughs is that there is a book, Pawn Sacrifice!, by IM Timothy Taylor and he has charted all kinds of pawn sacrifices. All the same he has forgotten or missed the “impossible break” where at least two pawns can take the sacrificed pawn. He mentions two examples of this rare kind of break in connection with a king attack and line opening but makes no further investigation of it. In this collection of important positions Nos. 103, 197, 244, 256 and 272 contain such a surprising pawn break. Don’t forget it! It’s not enough, as Nimzowitsch wrote in his famous books, to trammel a pawn, it must be blockaded permanently to stop its lust to expand. Published as an eBook in the United Kingdom in 2020 by Batsford, an imprint of Pavilion Books Company Ltd. 43 Great Ormond Street London WC1N 3HZ Copyright © Batsford 2018, 2020 Text copyright © Thomas Engqvist 2018, 2020 The moral rights of the author have been asserted. All rights reserved. No part of this publication may be reproduced, stored in retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the copyright owner. ISBN 978-1-84994-575-2 A CIP catalogue record for this book is available from the British Library. This book can be ordered direct from the publisher at www.pavilionbooks.com, or try your local bookshop.
0
You can add this document to your study collection(s)
Sign in Available only to authorized usersYou can add this document to your saved list
Sign in Available only to authorized users(For complaints, use another form )