Author: KarinsDad
Date: 21:24:26 12/30/98
Go up one level in this thread
On December 30, 1998 at 20:58:25, Chris Moreton wrote: >I hope I am not being ignorant here although I fear that I am. When a position >is evaluated as being favourable for one side, this can, I think, only be judged >with regards to available knowledge. This available knowledge can only be >provable if there is a forced mate (which could be proved with a chess program) >or a known draw that has been proven by those in the know. The initial >position, for example is not and maybe will never be provable as favouring one >side as this would require that all moves in game be searched to the conclusion >of the game with the eventual losing side always playing the most resistant >(optimal) moves. In my opinion, I think that only positions where there is a >forced mate or a known draw can be confidently, provably evaluated correctly. I >think in all other positions, there is no known thing as an optimal move. > >Please forgive me if I have made a fool of myself! > >Chris Actually, I tend to agree with you for the most part. However, I believe that there are positions that are so favorable for one side or the other than any human chess player looking at them would instantly say, "Yes, that's a won position for white and with correct play from both sides (whatever that exactly means), white will win.". However, a position like this may take a lot of moves to win. An example might be where white is up a queen in a closed pawn position where it is hard to break through to the other side (this is probably not the best example, but work with me on it). With correct play, white could break through, but it might take 20 moves (40 ply) to do so and a computer may have a hard time searching down far enough to prove it. Nonetheless, white should win and after enough moves, the computer will catch up to the human and determine this to be so (yes, I know that the computer will realize that the one side is up a lot of material, but the search paths may reveal little in the way of improving the position). Getting back to the original question, there are no proofs except in endgame studies. There are theories which have been used to win games and even the USCF allows the adjudication of games using players of higher strength. There is chess logic, but there are no real beginning to end proofs. I must assume (granted, only my opinion) that like my example above, a favorable position must by definition be capable of being maintained. However, as humans and computers both make mistakes (i.e. are incapable of searching all paths to their conclusions and hence do not have the "best move" for all positions), it is impossible to prove my position on this. It is a postulate at best (and maybe only a theory). KarinsDad > >On December 30, 1998 at 20:45:26, Mario Petrilli wrote: > >>Is there such a thing as chess logic? Similar to mathemathics? Is there such a >>thing as a "PROOF" in chess? >> >>Suppose that a chess position is determined to be favorable to one player. If >>each player proceeds to make optimal (best) moves, is the favorability of the >>game maintained? Can it be proven?
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.