Roanoke Times Copyright (c) 1995, Landmark Communications, Inc. DATE: FRIDAY, December 9, 1994 TAG: 9501120005 SECTION: EDITORIAL PAGE: A-22 EDITION: METRO SOURCE: DATELINE: LENGTH: Medium
Such outcomes arise not when players' interests are completely opposite, but rather when some of their interests overlap.
Consider, for example, the now-classic Prisoner's Dilemma, whose origins lie in thinking generated by John G. Nash's work in proving the existence of "equilibrium points" in non-zero-sum games between two players.
Suppose you're a wrongdoer who has been arrested with your partner for a joint crime, and now isolated from him while awaiting trial. You, your partner and the authorities all know that without testimony from at least one of you, there is only enough evidence to convict both of you on a relatively minor charge with a one-year sentence.
So you and your partner are offered, separately, a deal. Turn state's evidence, and you'll go free while the other gets five years. If you both turn state's evidence, you'll each get two years. You both know the other has also been offered the deal, but you might decide your own strategy without knowing his.
From your viewpoint, the logic of betrayal is compelling. You have no influence over what your partner does, but it doesn't matter: Whatever he does, you're better off betraying him. If he also betrays you, it's two years for you instead of five. If he doesn't betray you, it's a suspended sentence for you instead of one year.
Of course, he logically can come to the same conclusion. This is a Nash equilibrium point: Both of you, given the other's strategy, made the right decision. Yet the outcome is worse for both of you - two years each in prison instead of one - than before the deal was offered, or had both remained silent.
The Prisoner's Dilemma is not about people's failing to act rationally (though of course in real life they often do). In this case, both you and your partner have adopted a rational strategy. Rather, it is about situations in which all players do adopt rational strategies - yet because they are adopted individually rather than cooperatively, all players end up losing.
Cooperation can seem, considered in isolation, poor strategy. But it is often the best one.
by CNB