A Simple Strategy for Life

Vic Stenger

 Social scientists have found an elementary game that serves as a useful metaphor for many of the types of decisions that individuals, groups, and even nations must regularly make in their dealings with others. This game is called the Prisoner's Dilemma.

In the Prisoner's Dilemma, two players must decide whether or not to cooperate with each other in a series of moves. They are unaware of the other's choice in any given move, but they know the past history of both players' moves. An advantage exists in cooperating, as long as both do so. When both players cooperate in a given move, they each get 3 points. If both fail to cooperate, one point is deducted from each score. Now, here's the rub: Momentary advantage occurs when a player fails to cooperate in a given move in which the opponent cooperates. In that case, the non-cooperator gets 5 points while the gullible cooperator gets zero.

In playing this game, one obviously can't be a saint and always cooperate. The other guy will soon learn that, and win by non-cooperating. Similarly, the sociopath strategy of constant non-cooperation only works when your opponent is a saint. A non-saintly opponent quickly fights the sociopath to a draw.
Experiments with individuals playing the Prisoner's Dilemma have found no strategy superior to the very elementary one of tit-for-tat. In tit-for tat, you simply cooperate on a given move when your opponent cooperated on the previous move, and don't cooperate when your opponent failed to do so. While this will not win every game, it wins more than any other strategy that people have been able to dream up.

Note I said people. When computers using what are called genetic algorithms play the game, they are occasionally able to come up with complex strategies that are almost as good as tit-for-tat. Genetic algorithms are unprogrammed programs that start out with random operations but, by means of a kind of sexual mixing of internal parameters, are able to reproduce new programs, using natural selection to evolve algorithms that are better equipped to solve the desired problem. The algorithm that results develops, like life, as an efficient but non-unique solution to a given predicament. (Those who still doubt that purely material systems can spontaneously develop life and intelligence should read about genetic algorithms and the other remarkable developments in the budding field of Artificial Life ).

Though genetic algorithms are occasionally better at solving the Prisoner's Dilemma than tit-for-tat, tit-for-tat still wins most of the time. Furthermore, tit-for-tat also evolves as the best overall, genetically determined algorithm. That is, if people had never thought of tit-for-tat, the computer would have thought of it for them.

After learning about all this, I personally began applying tit-for-tat in my dealings with people who have given me trouble in the past. It works! (I use the sociopath strategy for saints, of course, but I rarely interact with Mother Theresa.)

The need for a strategy beyond sainthood or sociopathood usually arises in the work place, where you must collaborate with colleagues to advance your common goals and those of your employer. Often your colleagues and employers have additional agenda that do not always include you. Tit-for-tat. Your colleague shares workload with you and gives you credit for your contribution; do the same for your colleague next time. Your colleague drops a load on you, and hogs the credit; find a way to do the same in return. That colleague will find herself losing ground unless she starts using a little tit-for-tat herself.

Try it. I do not suggest that all decisions in life can be reduced to such a simple rule, but I have found that tit-for-tat provides a nice first step in figuring out the best way to respond to the actions of others.

Now, of course, this is far from Jesus's turn-the-other-cheek. But judging from events in Ireland and Bosnia, and the history of Christendom, Jesus's followers have only rarely practiced sainthood as a strategy for life. They know it doesn't work.

To learn more about genetic algorithms, the Prisoner's Dilemma, and other wonders, read Artificial Life by Steven Levy (Pantheon, 1992).