Optimal Strategies in Perfect-Information Stochastic Games with Tail Winning Conditions

We prove that optimal strategies exist in every perfect-information stochastic game with finitely many states and actions and a tail winning condition.