Assume you are playing a fivepiece endgame in chess Many mod

Assume you are playing a five-piece endgame in chess. Many modern chess programs contain built-in game trees for all endgames with limited number of pieces (such as five-piece endgames). Assume leaves of this tree are labeled by -1 if the position is a losing position, +1 if it is a winning position and 0 if it is a tie. Is algorithm alpha-beta pruning guaranteed find a winning strategy whenever it exists? Why?

Solution

Most chess computers try to match a possible end game to the game in progress, which is essentially a dynamic programming trace back Again, the endgame in question is avoidable though. looks like I ruffled some feathers here.

Thinking about it again, it seems like there is no theoretical problem with solving a finite game like chess. I would argue that chess is a bit more complicated than checkers in that a win is not necessarily by numerical exhaustion of pieces, but by a mate. My original assertion is probably wrong, but then again I think I\'ve pointed out something that is not yet satisfactorily proven .

I guess my thought experiment was that whenever a branch in the tree is taken, then the algorithm (or memorized paths) must find a path to a mate (without getting mated) for any possible branch on the opponent moves. After the discussion, I will buy that given more memory than we can possibly dream of, all these paths could be found.