Author: Rolf Tueschen
Date: 08:17:45 05/20/02
Go up one level in this thread
On May 20, 2002 at 10:30:32, Roy Eassa wrote: >On May 20, 2002 at 10:15:44, Rolf Tueschen wrote: > >>I would expect that the top >>programs would at best get Elo performances of 2200 - 2350 > > >That's just way too low by even the most pessimistic estimates I've ever heard. The reason for higher estimates is the unreflected handling of mere numbers. In my opinion the creative power of human beings is mostly estimated too pessimistically. In other words with some exceptions the _real_ play/fight human chessplayers vs computerchess programs has not yet begun. Would you doubt that the players have at least the same power as the programmers and book doctors? But for peanuts they don't change their chess. A World Champion won't even change for a million dollars! You know, it's not a question of a week's hard working or of one player. I'm talking about a new form of chess especially designed to play machines. What we saw until now is some basical of that 'chess'. What Eduard Nemeth is trying to do for example. But he with Elo 2100 basically tried to exploit openingstuff and when it doesn't work he might lose the game. I'm talking about calculating exactitude. The moment new chess (for playing machines) has been developped players like Eduard could train on tricks for the overall game. That's not magic, that's simply man's mind. And the ever actual solidity of the machine's horizon makes it very easy to be exploitated. Very trivial. Rolf Tueschen
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.