The branch of game theory that studies the interaction of non‐rational beings such as animals, or humans whose behaviour evolves under the influence of the environment in which they find themselves. At first sight, evolutionary game theory seems to contradict the basic premiss of game theory, namely that it formalizes the interdependence of rational actors. But actors may reach Nash equilibrium by evolution as well as by reasoning. If they are human, they may notice that some strategies work and others do not, without understanding why the strategies that work do so. Or they may read successful business books at airports. If they are animals, then natural selection may select for successful strategies. Animals that ‘play’ (actually, that are hard‐wired to act in accordance with) successful strategies survive to have more offspring than those that ‘play’ others. Therefore the gene for the successful strategy spreads in the population. A version of Nash equilibrium for evolutionary games is an evolutionarily stable strategy (ESS), a term invented by J. Maynard Smith in 1973 and popularized by Richard Dawkins, Robert Axelrod, and others. Ironically, an ESS may itself be unstable, because what is evolutionarily stable in a given population depends on the make‐up of the rest of the population. If most of the rest of the population are aggressive ‘hawks’, it pays to be a ‘dove’; if most of the rest of the population are pacifistic ‘doves’ (doves in the literary conception, not the biological one—doves are actually aggressive birds), it pays to be a hawk.