Outline Motivation and Modeling Philosophy Empirical Alternative I: Model of Cognitive Hierarchy (Camerer, Ho, and Chong, QJE, 2004) Empirical Alternative II: Quantal Response Equilibrium (McKelvey and Palfrey, GEB, 1995) Empirical Alternative III: Model of Noisy Introspection (Goeree and Holt, AER, 2001)
Modeling Principles PrincipleNash CH QRE NI Strategic Thinking Best Response Mutual Consistency
Quantal Response Equilibrium Player i choosing strategy j with probability ij in equilibrium. x ij is i’s expected payoff of choosing j given that others choose their strategy according to equilibrium profile . The values of depends on.
Example 1 q (1-q) p (1-p)
Nash Equilibrium q (1-q) p (1-p) q = (1- q) q = (1-p) = p p = 0.5
Example 1: =2
Prediction as a Function of
Example 2
QRE Prediction: A = B = 5
QRE Prediction: A = B = 100
Lieberman’s 2-person Zero-sum Game
MLE Estimation
Best-fitted QRE Model
QRE Plots
O’Neill’s 2-person Zero-sum Game
QRE and its Special Cases
QRE Estimates Across Time
QRE Plots
Rapoport and Boebel’s 2-person Zero-sum Game
QRE Estimates; Experiment I
QRE Estimates: Experiment II
QRE Estimates Across Time
QRE Plots
Ochs’s 2-person Non-zero Sum Games
QRE Estimates
QRE Plots
QRE Estimates
Prediction
QRE Plots
QRE versus CH
Summary