Experience-Weighted Attraction Learning in Games: Estimates From Weak-Link Games

Abstract

How does an equilibrium arise in a game? For decades, the implicit answer to this question was that players reasoned their way to an equilibrium, or adapted and evolved toward it in some unspecified way. Theorists have become interested in the specific details of how adaptation and evolution work. Much of this interest revolves around models in which players change their strategies or learn, and what equilibria might result under various learning rules. Our research is motivated by a different question: Which learning models describe human behavior best? This chapter proposes a general experience-weighed attraction (EWA) model and estimates the model parametrically using a small set of experimental data.