A characterization of sequential equilibrium strategies in infinitely repeated incomplete information games

It is shown that the set of sequential equilibria of an infinitely repeated incomplete information game coincides with a family of markov chains (with a state space of player type distributions and payoffs). Corresponding to each markov chain in the family there is a sequential equilibrium strategy pair, characterized as a pair of functions from the state space to the action spaces. These functions may be chosen independent of the given markov chain. In an equilibrium of this form, players at any point in time condition their actions only on the current value of the state variable.