Axelrod-Python
Give players probabilistic knowledge of length of game
set_player_information.rst describes how to give players in a match knowledge of the length of a particular game. I think it'd be cool to have an option to have that length be probabilistic, rather than either known or not known. Would be curious to see what kind of effects this has on optimal strategies. I also think a probabilistic estimate of game length more closely models real life behavior. Thoughts on this? Does this feature already exist somewhere and I'm just not seeing it?
How would this be implemented? IIUC players would need to know that a probabilistic end is possible and take that into consideration, which none currently are coded to do so.
We do have the ProbabilisticTournament variant, where players don't know the length of the game but it is drawn beforehand and fixed (just not made available to the players).
I was understanding a version of the "standard" tournament where players DO know the length of the match but that length is not actually correct. Essentially add noise to what the players know about the tournament.
If a match is going to have 50 turns, perhaps what is told to the players is sampled (for each match) from some distribution around 50.