Consumption war

The war of attrition is a model of competition in which rivals bid time rather than inflict injury: each maintains a display until one backs down, and only the one who remains gets the resource. Its central result is that neither a fixed timetable nor truthful signals of resolve can remain secure against exploitation. The stable strategy is to persist for an unpredictable length of time, on average matching the resource’s value, and to keep the moment of surrender unreadable. The model’s reach extends beyond animal contests: it illuminates how rules can reshape apparent conflicts, why cooperation may emerge through repeated encounters, and how biological outcomes depend on whose future reproduction is at stake.

The contest and its currency

In the version considered by John Maynard Smith, the contestants belong to a species in which dangerous combat is absent or injury is unlikely, perhaps because the animals are well armoured. Disputes are settled through conventional posturing. Each rival stands its ground and glares; eventually one turns away. The resource at stake may be valuable, but it is not infinitely valuable, and time spent threatening cannot be devoted to other necessary activities. Time therefore serves as the currency of a two-bidder auction.

The analogy to an ordinary auction has a crucial limit. In a sale, the winning bidder obtains the goods while the losing bidder need not pay the winning bid. In this contest, both rivals pay by spending time, but only one acquires the resource. That difference means a strategy that seems economically sensible when considered in isolation can fail once it meets an opponent following the same rule.

Why fixed limits fail

Suppose every animal could estimate exactly how much time a particular resource, such as a female, was worth and withdrew at that limit. A mutant willing to persist just a little longer would win against these fixed bidders, making their strategy vulnerable. Even if all contestants bid precisely the resource’s estimated value, they would withdraw together when the limit arrived, and neither would gain it. In that population, an individual that retreated immediately could avoid wasting time and spread its strategy.

Immediate retreat would not remain advantageous indefinitely. Once immediate retreaters became common, an animal prepared to wait a few seconds before giving up could win against them. Selection would then favour progressively longer persistence, until withdrawal times again approached the resource’s economic value. This verbal account suggests a cycle between extreme persistence and instant retreat, but the analysis described in the book rejects a simple oscillation as the result.

Instead, the evolutionarily stable strategy is a distribution of persistence times. Each individual continues for an unpredictable period, with the average duration corresponding to the true value of the resource. If a display is worth five minutes, a particular contest may last longer, less, or exactly five minutes. What matters is not that every rival stops at a precise five-minute mark, but that an opponent cannot know in advance how long the other will hold out on that occasion.

Concealment and the poker face

Uncertainty is strategically useful only while contestants do not disclose when they intend to quit. A reliable whisker-flicker that foretold withdrawal within a minute would give an opponent a simple advantage: wait out that minute. A rival nearing its own planned retreat could instead withdraw immediately, avoiding further wasted time. Natural selection would penalize whisker-flickering and comparable signs that reveal future behaviour. The poker face would evolve: surrender, when it comes, is sudden and unpredictable.

The model also explains why a clear signal of determination cannot simply become a stable substitute for concealment. If most animals raised their hackles only when they genuinely intended to persist, opponents could learn to withdraw immediately whenever they saw the display. Individuals with little intention of persisting could then imitate it and win cheaply. As deception spread, selection would favour opponents who called the bluff; liars would decline, and the cycle could reverse. Neither reliable truth-telling nor reliable lying is evolutionarily stable in this contest. Concealing the intended moment of retreat avoids giving an opponent a dependable rule to exploit.

Beyond the isolated contest

The account assumes a contest between opponents who are otherwise alike: equal in strength, weapons, armour, and the amount each has to gain. That symmetry makes the model tractable, but it is not a realistic description of every rivalry. Individuals can differ in size or fighting equipment, in the value of victory, or in what injury would cost them. An old male with little reproductive life remaining, for example, may have less to lose from injury than a younger rival. The model’s logic therefore supplies a baseline for considering how differences between contestants might alter the stable outcome, rather than a universal schedule for every contest.

The book places this analysis alongside the wider problem of strategic interaction. In the Prisoner’s Dilemma, two players simultaneously choose between cooperating and defecting, and the banker reveals the cards on which their winnings depend. Repeated play changes the situation: cooperation can be sustained when neither player knows when the final round will occur, leaving a long enough shadow of the future. By contrast, if the game is known to end after exactly one hundred rounds, backward reasoning makes defection the predictable choice on the last round, then on the ninety-ninth, and successively earlier. The comparison clarifies what the war of attrition shares with other strategic models: outcomes depend not only on what a resource is worth, but on information, timing, and the structure of interaction.

That structural point appears in human institutions and public events as well. A divorcing couple might benefit from treating separation as a nonzero-sum problem, especially to preserve family finances, but the professional rule described in the book permits a lawyer to represent only one member of the couple. The arrangement can push the two sides into an adversarial frame even when they would prefer an amicable settlement. In the 1977 Bristol City–Coventry City match, news that Sunderland had lost changed the incentives in the final minutes: a draw would keep both Bristol and Coventry in the First Division, and the players shifted from competing for goals to protecting the shared outcome. Like a contest’s rules about persistence and surrender, these examples show how the structure surrounding a choice can change what counts as winning.

Evolutionary interests and the limits of analogy

The same gene-centred reasoning that frames contests also informs the book’s account of parasites and hosts. A fluke may influence a snail to produce a thicker shell, prolonging the snail’s bodily survival while imposing an energetic and reproductive cost on the snail. The proposed explanation is that fluke genes can benefit from the host’s continued survival without sharing its particular interest in the host’s reproduction. The argument is presented as a conjecture that could be tested, not as an established experimental result.

Other cases sharpen the contrast. Nosema parasites in flour-beetle larvae produce a chemical analogue of juvenile hormone, preventing normal development into adults and leaving giant larvae more than twice the weight of a normal adult. Sacculina, a parasite of crabs, sends roots into the host’s tissues and castrates it, redirecting resources away from reproduction. In both examples, a change that serves parasite genes can come at the host’s reproductive expense. The proposed general test is whether parasite genes reach future generations through the same vehicles as host genes: when their transmission routes diverge, their interests may diverge; when they share an exit, cooperation becomes more likely.

The war of attrition is not itself a claim that rivals will cooperate. It describes how persistence, hidden intentions, and unequal costs can shape a contest over a valued resource. Set beside repeated games, institutional constraints, and host–parasite conflicts, it contributes to a larger argument: apparently simple behaviour cannot be judged by appearance alone. A display may conceal retreat, a rule may manufacture antagonism, and prolonged survival may benefit one set of genes while costing another its reproductive future.