Author: Dann Corbit
Date: 11:31:23 03/25/02
Go up one level in this thread
On March 25, 2002 at 14:02:47, Russell Reagan wrote: >I was thinking about what computer vs. human matches really mean. If Gulko >doesn't get a single win against the 4 computer programs, does that mean >anything? Does that mean that the computers are generally stronger? Or does it >mean that in the course of 8 games, he made 2 less than optimal moves? If the >latter, I don't believe that means anything as far as whether or not computers >are better than the best humans yet. Think of it as a data point in an experiment. By itself, it means very little. But when combined with many others we can draw more and more meaningful conclusions. >So Kasparov loses his last match against Deep Blue, Gulko will likely lose his >match against the computers, and what if Kramnik loses to Fritz? Does that >really mean anything? We still have a fairly small pool of games from world >class players vs. computers. You could even throw Gulko out of the "world class" >category, but someone else should make that decision, because I have no idea of >his playing level compared to a Kasparov of Kramnik. What we see is that there are different ways to play chess. You can do little tactical snacks where you edge out 1/100 of a pawn, or you can plan off into the distant future. Both methods win chess games and it might be that tactical snacking with inhuman perfection is stronger, in the long run. For sure, at some point it will be. The only question is "Are we there yet"? >It seems like it would take regular competition between world class human >players and computers for the consensus to be that computers are better than the >best human players. We all know computers are "really good", so unless we can >say with relative certainty that computers are better than the best human >players, do any of these matches mean anything? As far as I can tell, these >matches just lead to the conclusion that "we don't know". > >What do you think? I think it's a great idea, but it will be hard to get someone to pay for it and it will be even harder to get the GM's to do it unless the money is fantastic. Got anyone in mind to bankroll the project? The Junior team and the Rebel team have been doing efforts like this for many years on a small scale. Most of the relevant data we have is from these efforts.
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.