Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The truth about chess programs

Author: Tony Nichols

Date: 02:36:01 04/22/05

Go up one level in this thread


On April 22, 2005 at 05:29:32, George Tsavdaris wrote:

>On April 22, 2005 at 05:03:00, Odd Gunnar Malin wrote:
>>
>>In positional (or overall) play human are far above computers. It's easy to
>>check with first go through something like 'How good is your chess' by Danial
>>King or the guess the move lecture in Chessmaster. Then let the computer do the
>>same. I bet you as an 1800 player would score better than the comp, even I do it
>>in several games and I'm 1500. This messure the overall play and don't punish
>>you enough for your misses like in a normal game.
>
>I agree, not completely but i agree........
>
>>
>>Even in deep tactics a GM is better than a computer, the computers tactical
>>superior is in the speed to find short tactics (3-7 moves).
>
> Yes but deep tactics, need precise play and humans doesn't have this, so this
>is not an advantage for humans......
>
>>
>>So why do a computer win? It plays tacticaly at maybe 2400 level,
>
>No, it plays tactically (short tactics) in above 2900 level with absolute
>precise!
>
>>the difference
>>from a human is that it play at this level all the moves. A human player (even a
>>GM) don't play at his highest level the hole game. Just go throug a game with a
>>computer and you will se many small misses, and of course also many moves that
>>the comp initialy say is bad but when you walk through the variation it start
>>agreeing with the played move.
>>
>>Maybe you could say that a GM varies his play between 1800 and 2900 while the
>>computer play steady at 2400.
>
>Absolutely not! 1800??? No way.........
> I believe that they can vary between 2650 to 2900 and in very rare occasions
>they do a mistakes of a 2000 player, but this is too rare to be taken into
>consideration.....
>
>>
>>To conclude: A game of chess is a competition where the result is depended on
>>several factors, tactics, positional understanding, full consentration through
>>the hole game. And when you should judge a computers strengt messuered in either
>>Elo or title (GM/IM strength) you have to look only at the result of the
>>competition, this is what Elo and GM/IM norms are about.
>>
>
> The reason, for me, that computers can play at an equal level with top GM's is
>the amazing (short) tactical ability they have, the amazing defensive ability
>they have, but the most important is that they never make any blunders or even
>just mistakes that would cost the game and this means also that if they have a
>winning position they will win..........!

Hi, George
I don't agree that computers play short tactics above 2900 level. I think that
many test suites demonstrate that.
Regards
Tony



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.