Finite game move choices

[vb4]

I was just reading about competitive games where players have a finitie number of choices to make, such as chess. It was regarding the Nash Equilibrium scenario which proves that in finite games there is a state of equilibrium where neither player can do any better by changing strategy. Nash proved that there is always at least one such equilibrium for games with a finite number of strategic choices. Applying this to chess and evaluation etc I was thinking that in developing evaluators and etc to find and make that one perfect move may not be reachable and that we may have to accept the fact that regardless of how many parameters and complex our evaluators become that we may not be able to improve the suggested move at some point.

This Nash equilibrium was an interesting read and I was just thinking, since so much attention is given to finding that magical move, that perhaps we should also be exploring other new ways to improve the move selection process in a different way.

Just my 2 cents for what its worth,

Les

[Michael Sherwin]

Chess is complicated and is, at this point in time, not about finding perfect moves or a series of perfect moves. It is about finding consistently better moves than the opponent so that a winning advantage can be achived before the game fizzles to a draw. It is about taking advantage of the opponents mistakes due to an inferrior search. It is about not making gross errors of ones own due to a bad evaluator and just handing the win to the opponent. It is about the struggle over many moves, that may number in the hundreds. It is about holding on by a thread for 30, 40, 50 moves and eventually winning!