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.