= Evolutionarily stable strategy
{title2=$e(x^*,(1-\epsilon)x^*+\epsilon y)>e(y,(1-\epsilon)x^*+\epsilon y)$}
= ESS
{c}
{synonym}
A resident strategy is evolutionarily stable if every distinct mutant earns lower expected fitness when introduced at sufficiently small positive frequency. The upper bound on the frequency may depend on the mutant. In a finite symmetric matrix game, $e(x,y)=x^TAy$ is the expected payoff, with the second argument representing the population strategy. Affineness in this argument yields the <two-condition criterion for evolutionary stability>. Every evolutionarily stable strategy is a symmetric <Nash equilibrium>, but a tied best response must also be strictly outperformed against its own mutant population. The <Hawk-Dove game> supplies an interior stable mixture despite equality of all payoffs against the resident mixture.
Back to article page
