V_V comments on Botworld: a cellular automaton for studying self-modifying agents embedded in their environment - Less Wrong

50 Post author: So8res 12 April 2014 12:56AM

You are viewing a comment permalink. View the original post to see all comments and the full post content.

Comments (54)

You are viewing a single comment's thread. Show more comments above.

Comment author: V_V 11 April 2014 04:01:50PM 2 points [-]

For example There was no single winning strategy but kind of rock-paper-scissors types of programs.

That's true for any symmetric zero-sum game.

Comment author: Gunnar_Zarncke 14 April 2014 04:21:55PM 1 point [-]

Interesting I wasn't aware of that. Does it hold only for sufficiently complex games? Can you give a reference?

Comment author: V_V 14 April 2014 07:59:59PM *  1 point [-]

Hmm, if you were talking about "rock-paper-scissors" as an example of games without a pure Nash equilibrium, then some games have it and some don't. Intuitively, the more complex the game (the larger the number of pure strategies) the less likely that there is a pure Nash equilibrium, but that's not a strict rule.
However, under weak assumptions, there is always at least one, generally mixed, Nash equilibrium.

If by "winning strategy" you meant a mixed strategy that guarantees a greater than zero payoff, then in a two-player symmetric zero-sum game it can't exist:
if it existed both player would use it at the Nash equilibrium, and the sum of their expected payoffs would be greater than zero.