Newsreel |
|
|
Player A |
Player B |
|
| Both cooperate |
250 |
250 |
| Both defect |
50 |
50 |
| Defect/ Cooperate |
400 |
0 |
| Cooperate/ Defect |
0 |
400 |
As you can see, when both players cooperate, they both get a good reward. When both defect, they get much smaller rewards. However when one cooperates and the other defects, it's the 'good' player who loses out, getting no points at all, while the 'cheat' gets the biggest reward of all. Such is life!
Another factor is that players don't know how many rounds there will be in each game - it's chosen at random, and can be anything between six and 16. This makes choosing between long and short-term strategies much harder. That's a lot like life too.
Playing modes and game strategies
DNJ Dilemma comes with four built-in game-playing strategies (called
"Tit for Tat", "Random Defects", "Adaptive
Defects" and "Suspicious Tit for Tat"). You can play against
these strategies, either choosing which one the computer should use for each
game, or allowing the computer to choose one at random (in which case it
doesn't tell you which one it's been using until the game ends - this is the
'serious' way to play against the machine!)
You can also make the computer play against itself, either telling it which strategy to use for each player, or allowing it to choose its own. In this 'self-play' mode, you can tell the computer to play multiple games in a given strategy combination, for example 10 games of Tit for Tat against Adaptive Defects.
The computer strategies don't cheat, so never 'look' at the card you've played in the current round before playing its own, and don't 'look' at the number of rounds in the game either. However they will take into account the cards you've played in previous rounds, just as you're free to analyse their previous cards.
Copyright © 2001 Matt Publishing Limited. All rights reserved. No part of this site may be reproduced without the prior consent of the copyright holder.