Skip to content

Glossary · Game Theory

Prisoner's Dilemma

also: prisoners' dilemma · prisoner's dilemma game · iterated prisoner's dilemma · repeated prisoner's dilemma

Definition

The prisoner's dilemma is a two-player game in which defection is each player's dominant strategy, yet mutual defection leaves both worse off than mutual cooperation. Devised at RAND in 1950 by Flood and Dresher and named by Albert Tucker, it is the canonical model of price wars, promotional escalation, and any arms race where individual rationality produces collective loss.

The one-shot dilemma has a unique equilibrium, mutual defection, and no amount of communication fixes it because promises are not credible. The repeated version is where the interesting economics lives. Axelrod's 1984 tournaments showed that simple reciprocal strategies such as tit-for-tat sustain cooperation when the shadow of the future is long enough, and the folk theorem generalizes the point: with patient players, cooperation can be an equilibrium of the repeated game. Retail promotion, ad-auction bidding, and feature-parity races are dilemmas of this kind. The diagnostic questions are how often the game repeats, how visible defection is, how fast punishment can follow, and how heavily the players discount the future, because those parameters, not goodwill, determine whether a truce holds.

Essays on this concept