Why We Need Evolutionary Game Theory

Abstract

Abstract The original game-theoretic models and their extensions discussed in this book potentially shed light on a tremendous range of social behaviors. Despite the work outlined in this volume, surprisingly few game-theoretic models (other than sex-ratio models) have been rigorously empirically tested during the more than two decades that have elapsed since game theory became a part of evolutionary biology. The scarcity of explicit tests of game-theoretic models contrasts markedly with the extensive empirical testing of frequency-independent, fitness-maximization models (such as most optimal foraging models). Why is this? One possibility is that such models are difficult to test—for example, because it is difficult to demonstrate “equality of fitnesses” (or “equality of marginal fitness returns”) for discrete behavioral variants selectively maintained in a stable polymorphism or as parts of a mixed evolutionarily stable strategy (ESS). However, the latter answer isn’t satisfying for two reasons: First, frequency-independent optimization models (henceforth called “FIO” models), like game-theoretic models (henceforth called “GT” models), make predictions about, and thus can be tested by comparing, the relative fitnesses of alternative behavioral variants (e.g., Gomulkiewicz, this volume).