Repeated games tend to have large sets of equilibria. We also know that in the repeated prisoners dilemma there is a profusion of neutrally stable strategies, but no strategy that is evolutionarily stable. This paper shows that for all of these neutrally stable strategies there is a stepping...