The idea of applying game theory to problems in biology was given a formal basis nearly 50 years ago. Since then, the theory has advanced, and there have been numerous applications of it to a diversity of empirical systems. Most of this work takes a straightforward functional approach, finding a behavioral strategy that is evolutionarily stable in a well-specified specific situation. Relatively little attention has been devoted to the role of phylogeny, the role of learning during development, and the limitations imposed by the psychological and physiological mechanisms that bring about behavior in a complex world. Here I argue that a focus on these elements can improve the link between the theory and empirical systems and hence help us to understand how natural selection has shaped observed behavior.