This paper presents results from experiments with finitely repeated games with complete and incomplete information. We use two treatment variables: the number of rounds the game is played and the value of the probability that reflects the presence of incomplete information.