The prisoner's dilemma is a fun little game theory problem. Now you have the chance to play it against a computer opponent!
The prisoner's dilemma was originally formulated by mathematician Albert W. Tucker and has since become the classic example of a "non-zero sum" game in economics, political science, evolutionary biology, and of course game theory.

A "zero sum" game is simply a win-lose game such as tic-tac-toe. For every winner, there's a loser. If I win, you lose. Non-zero sum games allow for cooperation. There are moves that benefit both players, and this is what makes these games interesting.

In the prisoner's dilemma, you and Albert are picked up by the police and interrogated in separate cells without a chance to communicate with each other. For the purpose of this game, it makes no difference whether or not you or Albert actually committed the crime. You are both told the same thing:

If you both confess, you will both get four years in prison.
If neither of you confesses, the police will be able to pin part of the crime on you, and you'll both get two years.
If one of you confesses but the other doesn't, the confessor will make a deal with the police and will go free while the other one goes to jail for five years.
At first glance the correct strategy appears obvious. No matter what Albert does, you'll be better off "defecting" (confessing). Maddeningly, Albert realizes this as well, so you both end up getting four years. Ironically, if you had both "cooperated" (refused to confess), you would both be much better off.

And so the game becomes much more complicated than it first appeared. If you play repeatedly, the goal is to figure out Albert's strategy and use it to minimize your total jail time. Albert will be doing the same. Remember, the object of the game is not to screw Albert over. The object is to minimize your jail time. If this means ruthlessly exploiting Albert's generosity, then do so. If this means helping Albert out by cooperating, then do so.

To make this game more fun, I've given Albert several different strategies that were inspired by a chapter in Carl Sagan's book, Billions And Billions:

The Golden Rule - "Do unto others as you would have them do unto you." Albert always cooperates (doesn't confess). It's quite easy to take advantage of this innocent "turn the cheek" strategy.
The Brazen Rule - "Do unto others as they do unto you." Albert begins with a cautious defection (he confesses), but after that he does whatever you did last. A similar strategy which begins with cooperation is usually called "tit-for-tat."
The Brazen Rule 3 - Almost the same as the Brazen Rule. The exception is that Albert is a little more forgiving. If you defect (confess), Albert will forgive you about once every three times and cooperate the next time anyway.
The Iron Rule - "Do unto others as you wish, before they do it unto you." Albert always defects. Both of you tend to accumulate a large prison sentence.
??? - Albert decides randomly which of the above four strategies to use, and you have to figure out for yourself which one he's chosen. Albert does not randomly choose "confess" or "don't confess." Instead, he randomly chooses one of the above strategies and sticks with that one strategy until you change his strategy to something else.