• Choose a category
• All
• Classic-Fiction

Here is the evolutionary game theorist’s explanation, and it’s just what Darwinian theory needs to show that niceness can enhance fitness and thus get selected for. Fortunately for humans, in addition to imposing a small number of single “one-shot” games on our ancestors, nature also imposed on them (and still imposes on us) other quite different strategic choice problems. These problems are made up of a large number of repetitions of the same simple game played over and over with different people. Nowadays, you and a lot of other people go to your local convenience store and purchase something from a number of different people behind the counter. Each transaction is one round in an iterated PD game you and everyone else—shopper and salesperson—is playing. Similarly, in the Pleistocene, scavenging repeatedly presented our ancestors with the same repeated strategic choice problems over and over again. That’s what tribal life is all about.

In an iterated PD, there are many possible strategies players can adopt: always defect; always cooperate; cooperate on even-numbered games, defect on odd-numbered games; cooperate five times in a row, then defect five times in a row; cooperate until the other player defects, defect thereafter; defect until the other guy cooperates, then cooperate thereafter. The number of possible strategies is vast. Are there better strategies than always to defect? Is there a best strategy?

Under fairly common circumstances, there are several far better strategies in the iterated PD game than “always defect.” In iterated PD, always taking advantage of the other guy is almost never the dominant strategy. In many of these iterated PDs, the best strategy is a slightly more complicated strategy called “tit for tat.” Start out cooperating in round 1; then, in round 2, do whatever your opponent did in the previous round. Round 1 you cooperate. If the other player also cooperates, the next time you face that same player, cooperate again. If the other player cooperates in round 2, you know what to do in round 3. If in round 1 the other player defects, trying to free-ride on you, then in round 2 against that player, you have to defect as well. Following this rule after round 1, you will have to continue cooperating until the other guy defects, and vice versa. If you have both been cooperating through 33 rounds and in his 34th game against you the other player suddenly switches to defect, trying to free-ride on your cooperation, then in round 35 against him, you have to defect. If in round 35, your opponent switches back to cooperate, then in round 36, you should go back to cooperation. That’s why the strategy is called tit for tat.

Why is tit for tat almost always a better strategy in iterated PD than always defect? In a competition in which tit-for-tat players face each other, they cooperate all the time and so accumulate the second best payoff every time. That’s usually a lot better outcome than playing always defect: you get the successful free-rider’s top payoff the first time and only the third best payoff all the rest of the time. The more rounds in the iterated PD, and the more chances you have to play with tit-for-tat players and get the second highest payoff every time, the better the outcome for cooperating and the worse for defecting. Provided that the number of individual rounds in the iterated game is large enough, that the chances of playing against the same person more than once are high enough, and that the long-term payoff to cooperating in future games against other cooperators is high enough, it’s always best to play tit for tat, especially if you are relentlessly looking out only for number one.

It’s easy to set up one of these iterated PD games either among people or in a computer simulation. Give each of two test subjects two cards, one with a D for defect, the other with a C for cooperate. Each subject plays one card at the same time. If both play D, each gets $2; if both