REVISED

Z04_REND6289_10_IM_MOD4.QXD 5/2/08 3:01 PM Page 282

M O D U L E

Game Theory

TEACHING SUGGESTIONS
Teaching Suggestion M4.1: Game Theory and Conict. This module covers zero-sum, two-person games. Conict is a part of our world. Students can be asked to discuss the use of conict analysis and game theory in corporate and political settings. Teaching Suggestion M4.2: Use of Pure Strategy Games and Dominance. The use of pure strategy games and dominance shows students that some strategies or alternatives can be eliminated from consideration by carefully analyzing the situation. This is an important concept that can be applied to many problems in addition to game theory.

organizations, governments, and so on. A zero-sum game means that when one person wins the other person must lose. Therefore, the sum of gains and losses for both players will always be equal to zero because when one player wins, the other player loses. M4-2. The value of the game can be computed by multiplying the percentage that each player plays a given strategy times the game outcomes embodied in the table of the game. Since the optimal strategies for each player are obtained by equating the expected gains of both strategies for each player, there is a shortcut method for determining the value of the game. This shortcut method involves multiplying game outcomes times their probabilities of occurrence for any row or any column. M4-3. A pure strategy is one in which a player will always play one strategy in the game. Dominance can be used in game theory to reduce the size of the game. This is done by eliminating strategies that would never be played by one of the players of the game. M4-4. A strategy is dominated if there is another strategy that has outcomes at least as good as the outcomes of this rst strategy. Whenever a strategy is dominated, it can be eliminated from consideration because it will never be selected. M4-5. A saddle point is found by nding the largest number in each column and the smallest number in each row. If a number in the table has both of these characteristics, it is a saddle point. M4-6. If a game has a saddle point, it will be a pure strategy game. If there is no saddle point, the game is a mixed strategy game. M4-7. A mixed game is one in which each player would play every strategy a given percent of the time. In other words, there is no pure strategy in a mixed game. A mixed game can be solved by equating a players expected winnings for one of the strategies with his or her expected winnings for the other opponents strategy. M4-8. Strategy for X X2 Strategy for Y Y1 Value of game 6 M4-9. As strategy A1 Bs strategy B1 Value of game 19 M4-10. Xs strategy: 86Q 36(1 Q) 42Q 106(1 Q), 35 Q 57 22 1Q 57

ALTERNATIVE EXAMPLE
Alternative Example M4.1: Melinda (person A) and Stanley (person B) are involved in a competitive situation. Both have two strategies (1 and 2) that they can play. A table showing the winnings is presented below.
STANLEY (B) Melinda (A) Strategy 1 Strategy 2 Strategy 1 10 2 Strategy 2 1 7

To solve this game, we determine the strategies for both players. We begin with Melinda (player A). The equations are as follows: For player AMelinda: 10Q 2(1 Q) 1Q 7(1 Q) 10Q 2 2Q 1Q 7 7Q 14Q 5; Q 5/14 0.357 strategy 1 for player AMelinda 1 Q 9/14; Q 0.643 strategy 2 for player AMelinda For player BStanley: 10P 1(1 P) 2P 7(1 P) 10P 1 P 2P 7 7P 14P 6; P 6/14 0.429 strategy 1 for player BStanley 1 P 8/14 0.571 strategy 2 for player BStanley

SOLUTIONS TO DISCUSSION QUESTIONS AND PROBLEMS

M4-1. A two-person game is one in which only two players can participate. These players could be people, companies, other

282

REVISED
Z04_REND6289_10_IM_MOD4.QXD 5/2/08 3:01 PM Page 283

MODULE 4

GAME THEORY

283

Ys strategy: 86P 42(1 P) 36P 106(1 P), 32 P 57 25 1P 57 35 22 Value of game 86 36 66.70 57 57 M4-11. 21Q 89(1 Q) 116Q 3(1 Q) 95 86 Q ,1Q 181 181 86 95 value of game 21 89 181 181 56.69 M4-12. A1: A selects $5 bill A2: A selects $10 bill B1: B selects $1 bill B2: A selects $20 bill a. A1 A2 B1 6 11 B2 25 30

Bs strategy: 6P 25(1 P) 11P 30(1 P), 55 P 72 17 1P 72 So strategies remain identical c.

31 41 11 1.32 72 72 Since game value is negative, Id rather be B.

Value of game 6

M4-14. The game can be reduced to a 2 2 game, since X would never play X1 or X4 since X stands to lose in every eventuality under those two strategies. Thus, the game is Y1 X2 X3 12 4 Y2 8 12

Xs strategy: 12Q 4 (1 Q) 8Q 12(1 Q) Q 2/3, 1 Q 1/3 Ys strategy: 12P 8 (1 P) 4P 12(1 P) P 1/3, 1 P 2/3 Value of game 12 2/3 4 1/3 9.33 M4-15. A1: Shoe Town does no advertising. A2: Shoe Town invests $15,000 in advertising. B1: Fancy Foot does nothing. B2: Fancy Foot invests $10,000 in advertising. B3: Fancy Foot invests $20,000 in advertising. a. A1 A2 B1 0 3 B2 2 1 B3 5 1

b. Strategy for A: 6Q 11(1 Q) 25Q 30(1 Q), 41 Q , 72 31 1Q 72 Strategy for B: 6P 25(1 P) 11P 30(1 P), 55 P , 72 17 1P 72 31 41 c. Value of game 6 11 1.32 72 72 Since game value is positive, Id rather be A. M4-13. a. A1 A2 b. B1 6 11 As strategy: B2 25 30

This particular problem has a saddle point with strategies A2 and B3 and game value of 1. M4-16. a. A1 A2 B1 0 3,000 B2 B3 2,000 5,000 1,000 1,000

Once again we have a saddle point at A2 and B3.

6Q 11(1 Q) 25Q 30(1 Q), 41 Q , 72 31 1Q 72

REVISED
Z04_REND6289_10_IM_MOD4.QXD 5/2/08 3:01 PM Page 284

284

MODULE 4

GAME THEORY

M4-17. The value of the game is 3.17. The optimal strategies for A and B can be computed along with the value of the game using QM for Windows. The results are presented below.

M4-20. The best strategy for Petroleum Research is to play strategy 14 all of the time. Petroleum Research can expect to get a return of $3 million from this approach. These results are summarized below.

Mixed Strategy For player A: Probability Probability Probability Probability Probability Probability of of of of of of strategy strategy strategy strategy strategy strategy 1 2 3 4 5 6 0.390 0.244 0.366 0.000 0.000 0.000 Mixed Strategy For player A: Probability Probability Probability Probability Probability Probability Probability Probability Probability Probability Probability Probability Probability Probability Probability of of of of of of of of of of of of of of of strategy strategy strategy strategy strategy strategy strategy strategy strategy strategy strategy strategy strategy strategy strategy 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 1.000 0.000

For player B: Probability of strategy Probability of strategy Probability of strategy Value for this game is 1 2 3 0.190 0.707 0.102 3.17

M4-18. Strategy X2 is dominated by both X1 and X3, so we may eliminate X2. When this is eliminated, we nd that strategy Y3 is dominated by both Y1 and Y2, so we can eliminate Y3. The value of the game is 6. M4-19. Y1 P X1 X2 Q 1Q 2 6 Y2 1P 4 10 Expected gain 2P 4(1 P) 6P 10(1 P)

For player B: Probability Probability Probability Probability Probability of of of of of strategy strategy strategy strategy strategy 1 2 3 4 5 0.000 0.615 0.385 0.000 0.000 3.00

Value for this game is

Expected gain 2Q 6(1 Q) 4Q 10(1 Q) To solve this as a mixed strategy game, the expected gain for each decision should be the same. Taking the expected gain for player X, we have 2P 4(1 P) 6P 10(1 P) 2P 4 4P 6P 10 10P 6P 4 10 4P 10P 14 P 1.4 Also, 1 P 1 1.4 0.4 Thus, for a mixed strategy game, the probability that X would play strategy 1 must be 1.4 which is greater than 1 (this strategy must be played 140% of the time). This is impossible. Therefore, if a person mistakenly solves a pure strategy game with the techniques of a mixed strategy game, the fraction (probability) for one of the players will be greater than one (or less than zero).