In this paper we prove the validity of the revelation principle for arbitration games and also establish that a large class of game forms can be represented as a Bayesian Statistician. This adds force and appeal to the concept of an arbitration game.