Problem set 2 Microeconomics 2 af Jonas Knudsen Question 1. Alice and Bob play a game given in strategic form as follows: 1/2 L C R T (2; 1) (3; 4) (5; 3) M (4; 3) (1; 2) (2; 1) B (3; 8) (0; 6) (1; 6) . a) Eliminate iteratively all strictly dominated strategies If we start from the side of player 1, we notice that strategy B is strictly dominated by M and a rational agent will therefore never play B, it is therefore eliminated. From player 2's perspective, we notice that strategy R is strictly dominated by C and is therefore eliminated. The simplified game is given by the following matrix. 1/2 L C T (2; 1) (3; 4) M (4; 3) (1; 2) . There is no longer any strictly dominated strategies. b) Find all pure- and mixed strategy Nash equilibria of this game. Is is noticed that if player 1 plays the strategy T the best response of player 2 will be to play C. If player 1 plays the strategy M the best response of player 2 will be to play strategy L. We quickly notice that both (T ; C) and (M ; L) is pure nash equilibria. Now for the mixed Nash equilibria. A mixed strategy is a nash equilibrium if and only if (p ~ ? )TA(q~? ) >(p ~ )TA(q~? ) 8p ~ 2 S ? T ? T ? ~ ~ (p ~ ) B(q ) >(p ~ ) B(q ) 8q ~ 2 S Where A; B is the payoff matrices for player 1 and player 2 respectively. 1 Therefore, consider the two mixed strategies (1 ¡ p; p) and (1 ¡ q; q) and the following system of equations. @ @ 1¡q ( 1 ¡ p p )A = 0 , ( (2+2 p) (1¡ q)+(3¡2 p) q) = 0 q @p @p ,¡4q + 2 = 0 , q = 1 2 @ @ 1¡ q ( 1 ¡ p p )B = 0 , ( (1+2 p) (1¡ q)+(4¡2 p) q) = 0 q @q @q ,¡4 p+3 = 0 , p = 3 4 Meaning player 1's mixed strategy is given by 1? = (1 ¡ p; p) = ¡1 1 strategy is given by (1 ¡ q; q) = 2 ; 2 ¡1 3 ; , 4 4 and player 2's mixed Question 2. Consider the following strategic game: 1/2 L C R T (2; 5) (20; 0) (12; ¡1) M (4,-5) (8; ¡2) (10; 4) B (0; 1) (16; 2) (14; ¡5) . Is there a mixed strategy Nash equilibrium in which Player 1 uses strategies T and M with nonzero probability and does not use strategy B, and in which Player 2 uses strategies L and C with non-zero probability and does not use strategy R? Explain your answer. Let p1; p2 2 (0; 1) E(T ; p1) = 2p1 + 20p2 + (1 ¡ p1 ¡ p2)12 = ¡10p1 + 8p2 + 12 E(M ; p2) = 4p1 + 8p2 + (1 ¡ p1 ¡ p2)10 = ¡6p1 ¡ 2p2 + 10 E(B; p2) = 0p1 + 16p2 + (1 ¡ p1 ¡ p2)14 = ¡14p1 + 2p2 + 14 He is indifferent between T and M if ¡10p1 + 8p2 + 12 = ¡6p1 ¡ 2p2 + 10 , 2p1 ¡ 5p2 = 1 (1) And he is indifferent between M and B if ¡6p1 ¡ 2p2 + 10 = ¡14p1 + 2p2 + 14 , 2p1 ¡ p2 = 1 We now have a system of two equations: 2 ¡5 2 ¡1 p1 p2 0 1 1 1 p1 2 ¡5 ¡1 1 = , = =@ 2 A 1 p2 2 ¡1 1 0 2 (2) ¡1 1 Meaning player 2 is indifferent between T ; M ; B when (p1; p2; 1 ¡ p1 ¡ p2)= 2 ; 0; 2 : Meaning that C is never played. From just this we can already conclude that there does not exist a such mixed Nash equilibrium as that in the question. If it is a nash equilibrium an agent must be indifferent between all of his strategies. In this example, for player 2, he is only indifferent between his strategies when he plays M with probability 0. Therefore, there does not exist such a Nash equlibrium in this game. 3