You are here

The Iterated Prisoner's Dilemma

As a single game theory scenario, the prisoner's dilemma is a thought experiment revealing the shortcomings of selfishness. But when treated as an iterative game, it gains a strategic dimension and becomes a powerful working model for understanding human cooperation.

The premise of the iterated prisoner's dilemma is that same two people have the option to cooperate or defect repeatedly. The goal is to accumulate the least amount of jail time over the course of the game. If we plug in the numbers from my June post, we get the following chart.

  Player B
Cooperate Defect
Player A Cooperate 0.25 | 0.25 5 | 0
Defect 0 | 5 1 | 1


The iterated prisoner's dilemma, also known as the peace-war game, has long been studied by psychologists trying to understand the evolution of cooperation. In the early 1980s, researcher Robert Axelrod ran a tournament designed to test and compare iterated prisoner's dilemma strategies. The winner was Anotol Rapoport's tit-for-tat strategy. Rapoport began each contest by cooperating in the first round, and then in subsequent rounds doing whatever his opponent had done in the previous round. The success of a simple formula over more complex calculations was a surprise to Axelrod and the research community, and subsequent studies have uncovered scenarios where tit-for-tat does not work. It's ineffective, for example, against a coin-flip strategy, where the person determines whether to cooperate or defect randomly. It's also ineffective against an always-defect strategy.

Applying this to real-world cooperation scenarios, it becomes clear that the tit-for-tat strategy fails against two types of opponents: those who make impulsive choices, and those who wish harm on others. With most people, it's a good strategy to trust until they give us a reason not to. This is not the case with those who are looking out only for their own self interest, and those whose behavior is grounded in irrationality. Sooner or later, these two types of people will give us a reason not to cooperate, and we can reply accordingly. But for the majority of the people we encounter regularly and work with over time, cooperation and trust have tangible benefits. Even though it appears at first glance that we can get ahead by protecting our own self-interest, this proves in the long run to be a losing strategy.

333 users have voted.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer