Adaptive Zero Determinant Strategy

[dashiellfryer]

My recent graduate Emmanuel Estrada and I made a class that adapts its four vector in response to payoff. The player maintains zero determinant bounds each round. In our tests this player eventually outperforms all the standard zero determinant strategies. We tested by using 2000 rounds instead of the standard 200. We look forward to any comments.

[marcharper]

Hi all, quick update on this one. There are really four possibly strategies here depending on how one wants to update the underlying parameters each round. It's also not clear how much to update each parameter, so I'd like to run some evolutionary algorithms to try to fit these hyperparameters.