Learning in extensive-form games I. Self-confirming equilibria

A group of individuals repeatedly plays a fixed extensive-form game, using past play to forecast future actions. Each (asymptotically) maximizes his own immediate expected payoff, believing that others' play corresponds to the historical frequencies of past play. Because players observe only the path of play in each round, they may not learn how others act in parts of the game tree that are not reached infinitely often. Hence, differences and correlations in beliefs about out-of-equilibrium actions may persist indefinitely. The stable points of these learning processes are self-confirming equilibria, a weaker solution concept than Nash equilibria. Journal of Economic Literature Classification Numbers: C72, D83.