Reconciling ecology and evolutionary game theory or ‘When not to think cooperation’

Evolutionary game theory (EGT)—overwhelmingly employed today for the study of cooperation in a variety of systems, from microbes to cancer and from insect to human societies—started with the seminal 1973 paper by John Maynard Smith and George Price [1], in which they probed the logic of limited war in animal conflict. If fighting was essential to get access to mates and territory, then why did fights rarely lead to serious injury? Maynard Smith and Price developed game theory to show that limited war can be selected at the individual level. Owing to the explanatory potential of this first paper, and enabled by the elegant and powerful machinery of the soon-to-be-developed replicator dynamics [2, 3], EGT took off at an accelerated pace and began to shape expectations across systems and scales. But, even as it expanded its reach from animals to microbes [4–8] and from microbes to cancer [9–11], the field did not revisit a fundamental assumption of that first paper, which subsequently got weaved into the very fabric of the framework—that individual differences in reproduction are determined only by payoff from the game (i.e. in isolation, all individuals, regardless of strategy, were assumed to have identical intrinsic growth rates). Here, we argue that this original assumption substantially limits the scope of EGT. But, because it is not explicitly presented as a caveat, predictions of EGT have been empirically tested broadly across real systems, where the intrinsic growth rates are generally not equal. That has, unsurprisingly, led to puzzling findings and contentious debates [7, 12–15]. Flagging the high potential for confusion to arise from applications of EGT to empirical systems that it is not designed to study and suggesting a way forward constitute our main motivation for this work. In the process, we reestablish a dialog with ecology that can be fruitful both ways, e.g., by providing a so-far-elusive explanation for how diverse ecological communities can assemble evolutionarily.