In chapter 12 of The Selfish Gene, Dawkins introduces the game Prisoner's Dilemma. These are the rules of the game: There are two players. Each has two cards, one that says COOPERATE and one that says DEFECT. A round of a game consists of each player putting down one of their cards face down without knowing what the other player is going to put. When both players have put their card down, the banker flips them both. There are only 4 possible outcomes since each player only has two cards. Outcome 1: We both play COOPERATE. The banker pays each of us $3oo. This is called the Reward for mutual cooperation. Outcome 2: We both play DEFECT. The banker fines each of us $10. This is called the Punishment for mutual deflection. Outcome 3: You play COOPERATE, and I play DEFECT. Because I have defected, the banker pays me $500, also called the Temptation to defect. Because you have cooperated, the banker fines you $100, also called the Sucker's fee. Outcome 4: This is just the opposite of outcome 3. You play DEFECT and I play COOPERATE.
After explaining the game, Dawkins goes on to talk about a certain simulated tournament where different strategies played against eachother and against itself. The winner was a strategy called Tit for Tat which does what the opponnent did on the last turn. Its first move, however, is always COOPERATE. When I read this, I started thinking that this really isn't the best strategy. To prove my point I went to Google and looked for a Prisoner's Dilemma game online. I played the computer by always defecting, and since the computer was a "nice" strategy, meaning it started off cooperating, I immediately had an advantage. It turned out that the computer was programmed with the Tit for Tat strategy so after I had the advantage it just went on to mirror everything I did and I won by a 5 point margin. (Instead of $300, you get 3 points. Instead of $500, you get 5 points. Instead of a $100 Sucker's fee, you get 0 points. Instead of a mutual deflection fee of $10, you get 1 point.) What I realized later as I read on, was that it didn't matter whether you win or lose, but that you have a high point total. Sure, with my Always Deflect strategy I could never lose, but other strategies might fare better in point total if they didn't have so many DEFLECT run ins where each gets only 1 point. Certainly if in real life these points mean offspring to animals, a nice strategy must be the one that wins out.
Suscribirse a:
Enviar comentarios (Atom)
No hay comentarios:
Publicar un comentario