Abstract

This paper proposes an interpretation of the experimental evidence on tacit coordination games involving randomly matched players provided by Van Huyck, Battalio and Beil (1990), based on the notion of stochastic stability. When the model is calibrated with the parameters chosen in the experiment, it predicts that every strict Nash equilibrium is stochastically stable; therefore, in the long run we should not observe the emergence of any particular pattern of behavior, as suggested by the experimental evidence. The model is also compatible with the experimental evidence provided by Goeree and Holt (2005).