The Analytic Narratives project is based mainly on the use of strategic game theory to analyze particular historical outcomes. Strategic game theory relies on a strong version of standard rational choice assumptions - actors are assumed to be instrumentally rational, have complete information, and have the ability to do complex calculations such as backward induction over several moves. We argue that strategic game theory is useful for some but not all historical analysis, because in some cases these micro-level assumptions are unrealistic. When rational choice assumptions are unrealistic, we suggest that scholars should use types of game theory that relax these assumptions, such as evolutionary game theory and behavioral game theory. Our goal is to broaden the Analytic Narratives project by outlining the conditions under which each of these three different types of game theory will be most useful by specifying the scope conditions of rational choice assumptions.