Game-theoretic analysis is a well-established part of the toolkit of economic analysis. In crucial respects, however, game theory has failed to deliver on its original promise of generating sharp predictions of behavior in situations where neoclassical microeconomics has little to say. Experience has shown that in most situations, it is possible to tell a game-theoretic story to fit almost any possible outcome. We argue that, in general, any individually rational outcome of an economic interaction may be supported as the Nash equilibrium of an appropriately chosen game, and that a wide range of these outcomes will have an economically reasonable interpretation. We consider possible attempts to salvage the original objectives of the game-theoretic research program. In at least some cases, information on institutional structures and observations of interactions between agents can be used to limit the set of strategies that may be considered reasonable