Author: Mike S.
Date: 21:04:46 06/12/03
Go up one level in this thread
On June 12, 2003 at 23:39:53, Russell Reagan wrote: >(...) Not to say humans will not win a game or two, but the longer the match >>the more advantage for the chess computer. > >That seems to be exactly the opposite of what every expert in the world thinks. >Longer time controls, advantage humans. I guess he meant *many games* in a long match (not longer time controls), because humans get tired. Speaking of the relative strength in general, it will depend upon if the GM wants to win, or strictly wants to draw. While comps can win against the top-10 GMs then and when, when they go for a win, I wonder if the same comps could acknowledge the elo achieved by that, when they would have to play against "mediocre" GMs who strictly go for the draw. IOW, when a comp makes even 45% "only" against GMs with 2700 average, partially because these Super GMs must at least in some games take some risk to go for a win in their Man vs. Machine matches, and therefore are going to loose then and when, this would mean this comp performs 2664 elo. But to acknowledge 2664 against an opposition with i.e. 2550, the comp would have to make 65% (!) against them. Which I doubt that it would be possible currently, if these GMs would go for the draw always. Maybe it can make more than 50%, but I guess not 65%. Just an assumption. - But IIRC, Prof. Hyatt has explained here, based on his big experience from Crafty vs. GM games on ICC, that when a GM wants to draw, he can do it anytime. At least that's how I remember what he wrote. So it probably makes a big difference, which category of GMs is the opposition, and what they aim for. Regards, M.Scheidl
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.