This paper presents a two-step identification argument for a large class of quasilinear utility trading games, imputing agents' values using revealed preference based on their choices from a convex menu of expected outcomes available in equilibrium. This generalizes many existing two-step...