Author: Robert Hyatt
Date: 21:59:58 06/21/99
Go up one level in this thread
On June 22, 1999 at 00:03:26, blass uri wrote: > >On June 21, 1999 at 14:21:38, Robert Hyatt wrote: > ><snipped> >>Never said otherwise. But in how many games have you seen Kasparov do that? >>Once in 100? So that of our 5 games the other day, one was a 100-1 shot? >>That was the only point... > >I think that computers help humans to do tactical blunders. >I think that sokolov was surprised by good moves of Fritz in the opening. > >I think that Fritz had advantage and a good chance to win even without the >blunder of sokolov. Then I don't understand... as when DB ripped Kasparov in game 6, Kasparov played poorly and the computer was lucky. Here the GM was OK and Fritz played great? While still in book? The two circumstances seem similar. In _both_ games the human blundered in the opening and lost quickly... > ><snipped> >>In some types of tactics, yes. IE open board, kings exposed, pieces all over >>the place. In other types of tactics, computers have no chance. IE who finds >>the Shirov Bh3 sac? Not hard to understand as a human... but very deep >>(although fairly forcing) from a human perspective. > >It is not hard to find after you see it but I think that most GM could not find >it practically in a tournament time control. > >I read in the newspaper that a room full with grandmasters and computers did not >predict the move of shirov so GM's also failed to see the right move before >shirov played it. > >It is more easy to find a good move when you know that there is a good move. > I would never argue that point... my point was that a human _can_ out-calculate a computer in many positions... And get torn apart in many others... IE the machines are _not_ invincible tactically yet. ><snipped> >> So computers are _not_ >>tactically supreme just yet... Even the Rebel/Rohde game showed that quite >>clearly as well... > >It only exposed tactical holes of Rebel > >Hiarcs7.32 has no problem to find Bg2 instead of f4 at tournament time control >and the same for other top programs > and most likely Hiarcs would have exposed another hole against that particular GM... IE in the game Ferret played the other day in the 40/2 match, the human simply saw 'deeper' in the right position, and another 'hole' showed up. :) > > ><snipped> >>I am _not_ talking about a single tactical blunder. The Shredder/Ferret game >>wasn't a "single poor move" game. It was a series of bad moves leading to a >>totally lost (at least according to three GM players we had commenting on ICC >>during the game) position. So a blunder is one thing, playing like an 1800 >>for many moves is something else entirely... > >I do not trust the evaluation of GM's >I want to know if they can win against computers with black from the position >after move 25 of the game. > >Uri
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.