Prisoners’ Dilemma Static games: Examples 1 Lectures in Game Theory Spring 2011, Collection of examples 1 25.01.2011 G.B. Asheim, ECON4240-ex1 1 Prisoners’ Dilemma as a collective good game 2 C D C 2, 2 0, 3 D 3, 0 1, 1 25.01.2011 2 G.B. Asheim, ECON4240-ex1 Pigs C : Contributi ng to a collective good Gain for each : 2 Cost for contributo r : 3 1 2 C D C 2, 2 0, 3 D 3, 0 1, 1 D S P D P 4, 2 2, 3 D 6, - 1 0, 0 Payoff for each if both contribute s : 1 2 2 - 3 2 Payoff if only self contribute s : 1 2 - 3 0 Payoff if only other contribute s : 1 2 3 25.01.2011 3 G.B. Asheim, ECON4240-ex1 What should the players do? 1 2 L C 25.01.2011 4 G.B. Asheim, ECON4240-ex1 What should the players do? 1 R 2 Frequency for 2 : 13 1 3 L C R Exp utility for 1 : U 20, 40 10, 30 80, 20 Fre - 2 U 20, 40 10, 30 80, 20 15.9 M 50, 30 10, 40 30, 30 quency 8 M 50, 30 10, 40 30, 30 18.2 4 D 30, 0 20, 10 40, 0 22.9 D 30, 0 20, 10 40, 0 for 1 : Exp utility for 2 : 22.6 30.0 20.0 1 Is strategy D strictly dominated? Is strategy D strictly dominated? Payoff for 1 if 2 chooses R Payoff for 1 if 2 chooses R 2 1 L R U 6, 3 0, 1 5 M 2, 1 5, 0 4 3, 2 3, 1 M 5 4 3 Mixtures off Mi M and U D D M Mixtures of M and U 2 1 U 1 2 3 4 25.01.2011 5 Payoff for 1 if 2 chooses L 6 7 L R 6, 3 0, 1 M 2, 1 5, 0 D 4, 2 3, 1 Indifference curve if 1 U puts prob 5/9 on L and 4/9 on R M 5 2 4 D 3 Mixtures of M and U 2 1 2 3 4 25.01.2011 5 6, 3 0, 1 M 2, 1 5, 0 D 4, 2 3, 1 2 3 4 25.01.2011 5 6 U 2 chooses L 8 G.B. Asheim, ECON4240-ex1 Selects an integer between 1 and 100. Write this number, with your name, on a slip of paper. Make your o r selections simultaneously sim ltaneo sl and independently. independentl The average of the your numbers is then computed. The one whose number is closest to 70 percent of this average wins 50 kroner. (If two or more tie, then the prize will be shared in equal proportions.) 6 U 2 chooses L G.B. Asheim, ECON4240-ex1 U Each of you: Payoff for 1 if 1 R 70 percent game Is strategy D a best response? 1 L Payoff for 1 if 1 G.B. Asheim, ECON4240-ex1 Payoff for 1 if 2 chooses R 2 D 3 2 1 1 9 25.01.2011 G.B. Asheim, ECON4240-ex1 10 Hawk-Dove/Chicken as an R&D game Hawk-Dove/Chicken H : Develop new technology . Cost for developer : 4 Gain if the other has not developed : 5 for self, - 1 for other 1 2 Gain if the other has developed : 3 for self, - 3 for other 2 1 H D H D H 0, 0 3, 1 H 0, 0 3, 1 D 1, 3 2, 2 D 1, 3 2, 2 Payoff for each if both develop : 2 5 - 4 - 3 2 - 1 3 - 4 0 Payoff for a sole developer : 2 5 - 4 3 Payoff if only the other develops : 2 - 1 1 25.01.2011 G.B. Asheim, ECON4240-ex1 11 25.01.2011 G.B. Asheim, ECON4240-ex1 12 2 Battle of the Sexes 1 2 Opera M Opera Movie 0, 0 Movie 0, 0 1, 2 O O Pareto Coordination Player 1’s best response fn 2 3 2, 1 25.01.2011 Player 2’s best response fn 1 1 3 G.B. Asheim, ECON4240-ex1 M 13 Pareto Coordination (a variant) 1 25.01.2011 2 A 2 A B A 2, 2 0, 0 B 0, 0 1, 1 25.01.2011 Stag Hunt 1 B H 2 Stag Hare A 2, 2 0, 3 2 Stag 5, 5 0, 4 B 3 , 2 1, 1 Hare 4, 0 4, 4 0 G.B. Asheim, ECON4240-ex1 15 A game of trust 14 G.B. Asheim, ECON4240-ex1 25.01.2011 Player 2’s best response fn Player 1’s best response fn 1 5 S S G.B. Asheim, ECON4240-ex1 H 1 5 16 A game of trust Each of you is to choose A or B indep. of each other. 2 1 A player who chooses A, receives 100 kr from UofO. Choose A Choose B 100, 110 Choose A 100, 100 100,-500 A player who chooses B, receives 200 kr from UofO if all others also choose B, but must pay 500 kr if at least one other player chooses A. Will a guarantee help? Choose B -110, 500,100 100 200, 200 Write A or B on a piece of paper! 25.01.2011 G.B. Asheim, ECON4240-ex1 17 25.01.2011 G.B. Asheim, ECON4240-ex1 18 3 Illustrating the best response property and Illustrating the best response property and best response completeness best response completeness 1 2 L C R U 0, 4 4, 0 5, 1 M 4, 0 0, 4 D 1, 5 1, 6 1 L C R U 0, 4 4, 0 5, 1 6, 1 M 4, 0 0, 4 6, 1 6, 6 D 1, 5 1, 6 6, 6 {D} {R} has the best resp. property, but is not best resp. complete. 25.01.2011 19 G.B. Asheim, ECON4240-ex1 Illustrating the best response property and 2 L C R U 0, 4 4, 0 5, 1 M 4, 0 0, 4 D 1, 5 1, 6 {U, M} {L, C} has the best resp. property and is best resp. complete. 25.01.2011 1 C R U 0, 4 4, 0 5, 3 6, 1 M 4, 0 0, 4 5, 3 6, 6 D 3, 5 3, 5 6, 6 21 G.B. Asheim, ECON4240-ex1 Matching Pennies T 2 H T 1, - 1 - 1, 1 T - 1, 1 1, - 1 H 25.01.2011 G.B. Asheim, ECON4240-ex1 H H {U, M} {L, C} has the best resp. property. 25.01.2011 Player 1’s best response fn Police 1 2 T Drivers Legal Illegal Monitoring - 1, 0 - 1, - 4 Not monitoring 0, 0 23 25.01.2011 22 G.B. Asheim, ECON4240-ex1 Monitoring (of traffic) Player 2’s best response fn 1 2 1 2 L {U} {L} neither has the best resp. property nor is best resp. complete. 25.01.2011 20 G.B. Asheim, ECON4240-ex1 Even in games with only one Nash equilibrium, rational players may not coordinate best response completeness 1 2 I Drivers’ best response fn Police’s best response fn 1 3 L N 1 5 M - 3, 1 G.B. Asheim, ECON4240-ex1 24 4 LUPI (lowest unique positive integer) game A homogeneous good market price Each of you: Selects a positive integer (“whole number”). 2 (q1, q2 ) p (Q )q2 Write this number on a slip of paper. a b( q1 q2 ) q2 Inverse demand fn.: p y Make yyour selections simultaneouslyy and independently. The one choosing the lowest integer that nobody else chooses wins 50 kr. p (Q ) a bQ "Testing Game Theory in the Field: Swedish LUPI Lottery Games" Robert Östling, Joseph Tao-yi Wang, Eileen Chou & Colin Camerer, American Economic Journal: Microeconomics, forthcoming. q1 25.01.2011 Simultaneous quantity setting: Cournot comp. Nash equilibrium: Are there quantities for the two firms so that no firm will regret its own quantity when told of the quantity of the other firm? 1' s best respo p nse fn : q1 (a bqq2 ) / 2b 2' s best respo nse fn : q2 ( a bq1 ) / 2b q1c q2c a / 3b Q c q1c q2c 2a / 3b p (Q c ) a / 3 1 (q1c , q2c ) 2 (q1c , q2c ) a 2 / 9b Firm 1 chooses quantity first (1 is the leader) Stackelberg comp. (cont.) 27 The leader' s problem : max a b( q1 q2 The follower’s best response fn p q1 a bq1 2b ) q1 max a bq1 q1 q1 Is it best to be a leader or a follower? Th follower' The f ll r' s problem pr bl m : max a b(q1 q2 ) q2 q2 FOC : a bq1 2bq2 0 1 2 25.01.2011 q2 25.01.2011 q2 a 60 b 1 q1c q2c 20 c Q 40 p (Q c ) 20 c c 1 (q1 , q2 ) 2 (q1c , q2c ) 400 q1m q2m 30 Q 30 p (Q m ) 30 1 (q1m , q2m ) p (Q s ) a / 4 m q1 2 ( q1s , q2s ) p (Q s ) q2s a 2 / 16b 29 28 G.B. Asheim, ECON4240-ex1 q1s 30 q2s 15 Q 45 p(Q (Q s ) 15 s s s 1 ( q1 , q2 ) 450 2 (q1 , q2s ) 225 q2s a bq2s / 2b a / 4b G.B. Asheim, ECON4240-ex1 q2 f 2 ( q1 ) ( a bq1 ) / 2b s Q 1 ( q1s , q2s ) p (Q s ) q1s a 2 / 8b q1 q2 A numerical example i e y1s a / 2b i.e., Q s q1s q2s 3a / 4b Firm 2' s best respo nse fn : Q Iso profit curve FOC : a / 2 bq 0 1 for the leader q1 Firm 2 chooses quantity after having observed firm 1’s quantity (2 is the follower) p q1 G.B. Asheim, ECON4240-ex1 26 Sequential quantity setting: Stackelberg comp. q1 25.01.2011 quantity G.B. Asheim, ECON4240-ex1 Firms 1 and 2 choose quantities simultaneously & independently. q2 Q q2 25.01.2011 G.B. Asheim, ECON4240-ex1 2 (q1m , q2m ) 900 30 5