We call a correspondence, defined on the set of mixed strategy profiles, a generalized best reply correspondence if it has (1) a product structure, is (2) upper semi-continuous, (3) always includes a best reply to any mixed strategy profile, and is (4) convex- and closed-valued. For each...

We define the logit dynamic for games with continuous strategy spaces and establish its fundamental properties, i.e. the existence, uniqueness and continuity of solutions. We apply the dynamic to the analysis of the Burdett and Judd (1983) model of price dispersion. Our objective is to assess...

We study population game dynamics under which each revising agent tests each of his strategies a fixed number of times, with each play of each strategy being against a newly drawn opponent, and chooses the strategy whose total payoff was highest. In the Centipede game, these best experienced...

A population of agents recurrently plays a two-strategy population game. When an agent receives a revision opportunity, he chooses a new strategy using a noisy best response rule that satisfies mild regularity conditions; best response with mutations, logit choice, and probit choice are all...

We prove that any deterministic evolutionary dynamic satisfying four mild requirements fails to eliminate strictly dominated strategies in some games. We also show that existing elimination results for evolutionary dynamics are not robust to small changes in the specifications of the dynamics....

We call a correspondence, defined on the set of mixed strategy pro les, a generalized best reply correspondence if it (1) has a product structure, (2) is upper hemi-continuous, (3) always includes a best reply to any mixed strategy pro le, and (4) is convex- and closed-valued. For each...

We consider a model of stochastic evolution under general noisy best response protocols, allowing the probabilities of suboptimal choices to depend on their payoff consequences. Our analysis focuses on behavior in the small noise double limit: we first take the noise level in agents' decisions...

In Young (1993, 1998) agents are recurrently matched to play a finite game and almost always play a myopic best reply to a frequency distribution based on a sample from the recent history of play. He proves that in a generic class of finite n-player games, as the mutation rate tends to zero,...

In this paper I define an evolutionary stability criterion for learning rules. Using Monte Carlo simulations, I then apply this criterion to a class of learning rules that can be represented by Camerer and Ho's (1999) model of learning. This class contains perturbed versions of reinforcement and...