Review Lecture

Review Lecture - Midterm Material •  Ch. 1 ­11...

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: Midterm Material •  Ch. 1 ­11 •  Problem Set 1 ­4 •  “Represen:ng Games” & “Analysis of Sta:c SeCngs” posted notes •  Lectures corresponding to posted notes Midterm Material •  Normal form, dominance, best response, iterated dominance, common knowledge, ra:onalizability, beliefs, expected payoﬀs, Nash equilibrium, eﬃciency •  Construct extensive form representa:on from narra:ve •  Convert extensive form to normal form •  Matrix representa:on: ﬁnding ra:onalizable strategies, pure and mixed strategy Nash equilibria •  Find ra:onalizable strategies and Nash equilibria in oligopoly, partnership, loca:on 1 What is NOT on the exam •  Congruity, congruous sets (pg. 90 ­91) •  Contracts (Ch. 13) Ch 7: #5 and #6 •  Suppose that in some two ­player game, s1 is a ra:onalizable strategy for player 1. If s2 is a best response to s1, is s2 a ra:onalizable strategy for player 2? •  Suppose that in some two ­player game, s1 is a ra:onalizable strategy for player 1. If s1 is a best response to s2, can you conclude that s2 is a ra:onalizable strategy for player 2? 2 Ch 7: #5 Suppose that in some two ­player game, s1 is a ra:onalizable strategy for player 1. If s2 is a best response to s1, is s2 a ra:onalizable strategy for player 2? R={U, M, D} x {L, R} 2 1 L U 1, 1 M 2, 1 D 1, 2 C 1, 0 2, 0 1, 0 R 2, 2 1, 1 2, 1 L is a best response to D. D may ra:onally be played. Thus L is ra:onalizable. Similarly, R is a best response to U. U may ra:onally be played. Thus R is ra:onalizable. Ch 7: #6 Suppose that in some two ­player game, s1 is a ra:onalizable strategy for player 1. If s1 is a best response to s2, can you conclude that s2 is a ra:onalizable strategy for player 2? R={U, M, D} x {L, R} 2 1 L U 1, 1 M 2, 1 D 1, 2 C 1, 0 2, 0 1, 0 R 2, 2 1, 1 2, 1 M is a best response to C, but we cannot conclude from that that the C is ra:onalizable. M is a best response to L, and just happens to also be best response to C, which will not be ra:onally played. 3 Ra:onalizability and Nash Equilibrium •  Every Nash equilibrium is in the set of ra:onalizable strategies. –  Nash equilibrium strategy proﬁle is always ra:onalizable. •  Every ra:onalizable strategy proﬁle is NOT a Nash equilibrium. •  Nash equilibrium is a reﬁnement of the ra:onalizable strategies. Ra:onalizability and Nash Equilibrium The set of ra:onalizable strategies have mutual consistency in terms of the en:re set of ra:onalizable strategies. But there may not be mutual consistency in the beliefs of a speciﬁc strategy proﬁle. A Nash equilibrium will have mutual consistency in the beliefs of that speciﬁc strategy proﬁle. 4 Ra:onalizability and Nash Equilibrium •  •  •  •  Communica:on Social ins:tu:ons Historical factors Culture … facilitate coordina:on of beliefs and behavior. •  In a Nash equilibrium, beliefs are coordinated Ra:onalizability •  Each player’s ra:onalizable strategies includes every best response to her opponents’ ra:onalizable strategies. •  Players may have doubts about what others will do. •  Players beliefs may not accurately describe the others actually do. 5 Solu:on Concepts •  Dominance –  Assump:on: Ra:onality –  Never play a strategy that is never a best response to anything. •  Ra:onalizability –  Assump:on: Common Knowledge of Ra:onality –  Never play a strategy that is only a best response to a strategy the other player will not ra:onally select, (with common knowledge of ra:onality). •  Nash Equilibrium –  Assump:on: Common Knowledge of Ra:onality, correct beliefs in equilibrium –  Never play a strategy that is only a best response to a strategy the other player does not actually select. Indiﬀerence as Key to MSNE •  In an MSNE, a mixed strategy is a best response for at least one player. •  If a mixed strategy is a best response, it must be that each of the pure strategies involved in the mix must itself be a best response. •  Each of the pure strategies involved in the mix must yield the same expected payoﬀ. 6 Indiﬀerence as Key to MSNE 1 2 (q) X (1 ­q) Y (p) A 2,3 6,0 (1 ­p) B 2,1 0,2 In an MSNE of this game, one player plays a pure strategy and the other mixes over his two strategies. Which player can play a pure strategy in an MSNE? Player 2 playing X. If a player is playing a mix, while the other is playing a pure, the player playing a mix must be indiﬀerent between at least two strategies. Strategic Tensions •  First: conﬂict b/w individual and group incen:ves –  Prisoner’s Dilemma •  Second: strategic uncertainty –  Stag Hunt •  Third: ineﬃcient coordina:on –  Pareto Coordina:on 7 Strategic Tensions •  First: conﬂict b/w individual and group incen:ves –  Prisoner’s Dilemma •  Second: strategic uncertainty –  Stag Hunt •  Third: ineﬃcient coordina:on –  Pareto Coordina:on Bertrand Model NE when calculus does not apply •  There are a small number of ﬁrms in the market. –  All ﬁrms have constant marginal costs. •  Firms produce a homogeneous good. –  Consumers choose the cheapest product. •  Firms choose price, –  Quan:ty, Q(p) is determined by the market demand func:on. 8 Bertrand (Price Compe::on) Example Example: QD(p) = 500 – p; c1 = c2 = 100 Firm 1’s proﬁt func:on: Best responses are not well ­deﬁned Example: QD(p) = 500 – p; c1 = c2 = 100 Suppose p2=\$200. \$199.99 is not the highest possible price lower than 100. \$399.99 for 2 ­> p1=\$199.995 \$599.99 for 3 ­> p1=\$199.9967 … Gas sta:ons also use frac:onal cents. There can always be a number smaller than epsilon. f p2>pm where pm , BR is pm 9 Best Responses in Bertrand Bertrand Model Nash equilibrium with 2 ﬁrms: •  If both ﬁrms have iden:cal marginal costs, c, then the only Nash equilibrium is that both ﬁrms set p = c 10 Best Response Func:ons Consider a two ­player game. S1 = [0, 1] and S2 = [0, 1]. The graph illustrates each player’s best response func:on. Which line is player 1’s best response func:on? (the solid or striped)? Which line is player 2’s best response func:on? (the solid or striped)? A best response func:on for Player 2 must specify a strategy for any belief about player 1. It must reach from the lex side to right side. A best response func:on for Player 1 must specify a strategy for any belief about player 2. It must reach from the boyom to the top. The dashed line could be a BR for either player as it reaches from boyom to top and lex to right. However, the solid line can only be a BR for player 2. It cannot be a BR func:on for player 1 because it does not specify a strategy for all beliefs of what player 2 may play. It does not reach from the boyom to the top. 11 ...
View Full Document

This document was uploaded on 11/30/2011.

Ask a homework question - tutors are online