This paper proposes and studies the rationalizable foresight dynamics. A normal form game is repeatedly played in a random matching fashion by a continuum of agents who make decisions at stochastic points in time. A rationalizable foresight path is a feasible path of action distribution along which each agent takes an action that maximizes his expected discounted payoff against another path which is in turn a rationalizable foresight path. We consider a set-valued stability concept under this dynamics and compare it with the corresponding concept under the perfect foresight dynamics.

Key Words: rationalizability, Nash equilibrium, rationalizable foresight, perfect foresight, stability under rationalizable foresight (RF-stability), stability under perfect foresight (PF-stability).

Games and Economic Behavior 56 (2006), 299-322.
First draft: February 26, 2002; this version: June 15, 2005. PDF file