A Constructive Proof that Learning in Repeated Games Leads to Nash Equilibria

Abstract

This paper extends the convergence result in Kalai and Lehrer (1993a, 1993b) to a class of games where players have a payoff function continuous for the product topology. Provided that 1) every player maximizes her expected payoff against her own beliefs, 2) every player updates her beliefs in a Bayesian manner, and 3) prior beliefs of other players' strategies have a grain of truth, we construct a Nash equilibrium such that, after some finite time, the equilibrium outcome of the above game is arbitrarily close to the constructed Nash equilibrium.