The Logic of Animal Conflict
Why don't animals always fight to the death over resources? The hawk-dove game, formulated by John Maynard Smith, provides an elegant answer using evolutionary game theory. By modeling two simple strategies — escalate (Hawk) or display and retreat (Dove) — the game reveals that mixed populations are the inevitable outcome of evolution when fighting carries significant costs.
Payoff Matrix and Dynamics
The game's structure is captured in a 2x2 payoff matrix. Hawk vs Hawk: both risk injury, expected payoff (V-C)/2. Hawk vs Dove: Hawk takes all, payoff V; Dove retreats, payoff 0. Dove vs Dove: they share after display, each gets V/2 minus display cost D. The replicator dynamics drive the population toward the ESS frequency where Hawk and Dove fitness are equal.
The ESS Mixed Strategy
When injury cost C exceeds resource value V, the ESS hawk frequency is p* = V/C. This can be interpreted two ways: each individual plays Hawk with probability V/C (mixed strategy), or the population consists of a fraction V/C of pure Hawks. At this equilibrium, no mutant strategy can invade — any deviation from p* reduces the mutant's fitness below the residents'.
Extensions and Reality
Real animal conflicts are richer than the basic hawk-dove model. The war of attrition adds duration to contests. Asymmetric games (e.g., owner vs intruder) produce conditional strategies like 'play Hawk if you're the owner.' Sequential assessment models allow contestants to update their estimate of winning probability during the fight. These extensions explain the nuanced fighting behaviors observed in nature.