Given their immense costs, extended interstate wars seem hard to explain rationally, and hence appear to be fertile ground for theories grounded in psychology. Most existing work on war duration, however, neglects psychology, and even when psychological biases are explicitly incorporated into theories, their implications typically simply exacerbate rationalist factors impeding settlement. I argue that three central difficulties complicate efforts to apply insights from psychology to explain war duration. First, many psychological biases produce empirically intractable predictions because core concepts cannot be operationalized clearly. Second, common psychological biases that might produce extended violence, such as sunk cost bias, do not produce good explanations for shorter conflicts. Third, in the few cases in which psychology produces compelling hypotheses, extant rationalist arguments point in the same direction.