The Iterated Prisoner's Dilemma Competition:
Celebrating the 20th Anniversary


Breaking News:

·         A book 'The Iterated Prisoners' Dilemma: 20 years on' published by the World Scientific


You and a friend have committed a crime and have been caught. You are being held in separate cells. You are both offered a deal but have to decide what to do. But you are not allowed to communicate with your partner and you will not be told what they have decided until you have made a decision.
Essentially the deal is this.

What do you do?

The dilemma you are faced with is that it is better for you to confess, UNLESS you both deny taking part in the crime. BUT, if you think that you partner will deny taking part, then you CONFESS and go free. See the dilemma?

To read another (better!) description of the prisoners dilemma see the Stanford Encyclopedia of Philosophy entry and, to play a game, visit

Another way of looking at the Prisoners Dilemma can be found here. This is the definition used in this competition.


The Competition

We are planning a series of competitions (the first was held at the Congress of Evolutionary Computation in July 2003, the second will be held in April 2005) to celebrate, pay tribute to and update the competitions run by Axelrod [AXE84]. Importantly we aim to provide the environment and data to allow researchers to conduct investigations into the latest developments surrounding the iterated prisoners dilemma. The competitions we originally held were 1-3, below. For the second set of competitions, we have introduced a new category (number 4, below).

  1. Re-run the original experiment of Axelrod. We aim to see if a tit-for-tat strategy still dominates or whether somebody can develop a better strategy taking into account that there has been 20 years since the original competition (for example, tit-for-2-tats was claimed to be better than tit-for-tat). In addition, we would expect many more entries that the 62 received by Axelrod (largely due to the internet), thus giving us access to much more data.
  2. Organise a competition that has noise in the data. That is, a signal to cooperate or defect could be mis-executed.
  3. Allow competitors to submit a strategy to an IPD that has more than one player and more than one payoff, that is, multi player and multi-choice.
  4. This competition will mirror the original competition of Axelrod. That is, the rules are

More details about the competition can be seen here. Please note that we wish to make this competition open to as many people as possible - not just the academic community.

Please note: If you submitted a strategy(s) to the first competition, they will automatically be entered for the 2nd competition. In addition, if you only entered a single strategy, this will automatically be entered into competition 4, see above

Important dates can be seen here.

Other Details

Page hits (since 7th November 2003) :