This paper provides a theoretical foundation for Markov (perfect) equilibria in repeated games with asynchronous moves that is based on memory costs. We show that if players incur a ``complexity cost'' which depends on the memory length required by their strategies, then any rationalizable...