Axelrod’s Tournament

[imankulov]

Goal

Learn about prisoner's dilemma and it's iterated variation. Try different strategies with Python's "axelrod" package. Explore Python implementation of them, and write your own strategy

References

• Prisoner's dilemma on Wikipedia
• Axelrod's tournament on Wikipedia
• axelrod Python package

[nolasconapoleao]

Proposition for Challenge:

Explore library

• Match a Tit-for-tat to a Grudger and play 20 rounds
• Match a Grudger to 2 DefectorHunters and play 50 rounds

Write a new strategy

• Strategy that decides randomly every 3 turns
• Strategy that uses 3 turns for its decision
• Strategy that does not use the history of the last round
• Strategy that outscores a grudger against a tit-for-tater (with noise)
• Strategy that forgives offenses after 2 rounds
• Strategy that can spot a tit-for-tatter and cooperates with it (no noise)

Analyse match results

• Change payoff matrix for game to a zero sum game
• Organize strategies by score
• Plot graphic of results vs strategies
• Visualize comparison between all strategies permutations (round-robin)
• Introduce noise in tournament

Creating matches: https://axelrod.readthedocs.io/en/stable/tutorials/getting_started/match.html Visualize results: https://axelrod.readthedocs.io/en/stable/tutorials/getting_started/visualising_results.html Create matches: https://axelrod.readthedocs.io/en/stable/tutorials/getting_started/match.html

[imankulov]

Great overview 😋 I wonder though how we can realistically fit this (especially the analyze part) into a two hours long schedule (de-facto, a bit more than one hour in fact). Should we aim at extending the meetup to 3-4 hours maybe?