Author: Walter Faxon
Date: 22:47:57 01/08/03
Go up one level in this thread
On January 08, 2003 at 05:28:02, Uri Blass wrote: >On January 08, 2003 at 02:58:07, Walter Faxon wrote: > >>On January 07, 2003 at 05:51:26, Graham Laight wrote: <snip> >>>It might suggest it, but it falls far short of proving it. If you go to the SSDF >>>website ( http://w1.859.telia.com/~u85924109/ssdf/ ) and look at the games, >>>you'll see that games between stronger chess programs are more likely to end in >>>draws than games between weaker ones. This, together with the fact that it has >>>been proven that there's no forced win of material in the first 30 moves, proves >>>beyond reasonable doubt that chess is a draw with optimal play. >> >> >>I disagree. Stronger computers draw each other more often because each can >>avoid falling victim to the other's tactics. They are not exploiting subtle >>advantages; they don't know how. > >I disagree. > >Computers have evaluation that is not only material. >They can win games by slightly increasing their positional advantage and not >because of tactics. > > > They stumble around a while and if neither >>side happens into the winning line that neither side sees, they draw. Why else >>is a computer plus a competent human player (not necessarily a grandmaster or >>even a master) so much stronger than a computer alone? > >Computer plus human may be stronger than a computer in long time control because >of the following 2 observations: > >1)Computers do not know to search the right lines so human can help the computer >to extend the right lines. > >2)There are positions that computers do not understand so in these positions >humans can use their common sense and play different moves then the move that >the computer suggest. > > >Computer+human is often weaker than computer without help. >If the human use the computer only for check blunder then I expect >computer+human(who is not better than a master) to be significantly weaker than >a computer without help. > >Uri Uri, I do understand that current programs have better evaluators than just "mate + material", and can often exploit lesser advantages. The question is a matter of degree. If realizing a particular advantage in a particular position by search alone requires a search of N ply (with the position at ply N-1 looking bad), then searching N-1 ply just doesn't cut it. And as our computers and programs get better this will remain true; just the value of N gets larger. We need a new approach, not just faster searchers and better evaluators. You agree that computer + (sufficiently strong) human is better than computer alone. The human is bringing something different to the table. Unless we can discover and program more of that, there will always be positions that are just "a ply too far" away, and farther. And we won't find that something different as long as computer chess competitions remain focused on pure chess performance. Fast searching, which is pretty well understood by now, will annihilate early-generation limited-search programs. We need to give the latter their own sandbox to play in, to allow what AI calls "competitive evolution". -- Walter
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.