Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Why comps play without knowledge of opponent?

Author: James Robertson

Date: 17:31:28 01/30/99

Go up one level in this thread


On January 30, 1999 at 17:52:47, Reynolds Takata wrote:

> The
>truth of the matter is that a Chess program beat an I.M(Dean Hergott) in a 40/2
>match, further there are other strong IM's that have been beaten similarly in
>matches just ask Amir Ban.  These defeats were carried out on mere P200s. We
>have P450s now(probably overclockable on top of it).  We have a a micro holding
>Anand 2700+ to a draw,

No; it was mashed to a pulp. Anand scored THREE TIMES as many points as Rebel
did!!!!!

>and almost holding him in the other game.  We have comps
>having 2600+ performance ratings  against GM's at 40/2 at Aegon(again on slower
>comps),

I don't believe there was a single computer that faced only GMs. And it is
possible to achieve a GM norm if you beat enough IMs.

>and lastly a reported Fritz in disguise winning a 40/2 tournament
>blowing away a GM.

Sokolov recently blew away Kasparov at Wijk aan Zee, and he is 150 points lower.
Freak results happen all the time, like J. Polgar's recent loss to a 2350.

>In fact In a 6 game match, I doubt there is a legit
>2500-2550 GM in the world, that would dominate any of the top 3 micros in 40/2.
>In fact the likelyhood is that a great many  GM's would lose the match probably
>in convincing fashion

Give your evidence for this. What I see above does not hold water....

>(especially with the provision, that the operator is
>allowed to prepare a specific opeining book against each opponent, just as GM's
>prepare for their own competition).   We have GM's and IM's both, saying that
>the latest comps are GM strength.  Of course there are a few dissenters(and that
>number is dropping), of those dissenters(GM's), i'm sure when they are asked how
>they would dominate the top comps running on P450s i'm sure the noise would be
>deafeningly silent(though there might be a few holdouts with some excuse).
>  Something that is quite often left out of these discussions is that GM's have
>the distinct advantage of knowing far, far more about their opponent when it is
>a comp as opposed to when it is a human.

If you program a computer to have nerves, fatigue, and off days I will take this
as a legitimate complaint. We put up with our inherent weaknesses, let them put
up with theirs.

>So if a comp loses, it doesn't
>necessarily mean it didn't play at GM strength it may simply mean that it was >at a greater disadvantage than most other players. This because the opponent >had more knowledge about the comp than he/she would have about a human >opponent.

Refer to my above comment.

>Kasparov himself made a arguement in direct relationship to this, only in
>reverse.  Kasparov argued that he was at a distinct disadvantage against Deep
>Blue, because he claimed that deep blue had been prepared for him.  This >eaning
>the comp knew good lines to play against him specifically, how to perhaps aim
>for types of positions, that in the past Kasp hasn't been best at etc.  While
>he had no information(or at least not much), about the strengths and weaknesses
>of Deep Blue.  He was playing against as he put it (Unknown Entity).  At the
>time that line fell upon mostly deaf ears.

The loser will say anything. Kasp obviously thought that it wasn't that
important when he agreed to play the match. :)

>Comps are in the same situation
>frequently there opponent knows about them, yet they know nothing of the
>opponent.  Thus don't taylor their play against the opponent.  This is called
>Gamesmanship.  When Kasparov loses a game it doesn't mean he wasn't playing GM
>strength.  All it means is that his opponent through pure chessic ideas and
>gamesmanship beat him.  When comps lose, it doesn't necessarily mean that they
>aren't playing GM strength either.  Comps are equal to GM's in overall ability
>(this of course not meaning that they are positionally equal, but when their
>tactical superiority is added to their positional strength, as a whole then
>indeed yes relatively equal an GM strength.

I don't follow your argument here...

>   So in pure chessic terms GM
>strength in Gamesmanship no.  Evem despite this lack of gamesmanship comps hold
>their own with GMs.  If you honestly think that a lesser GM(2500) playing Rebel
>10 in a 40/2 match say 6 games would win on his pure chessic abilities(meaning
>coming in without knowing the opponent, after all that's what the comp is doing.
> Then indeed you have been hypnotized

Look into my eyes........ :)

>by the rhetoric of people who want to save
>face rather than acknowldege the evidence.

I do not understand what you mean by "save face", since you present no evidence
that contradicts my views.

>
>R. Takata
>USCF Life Master
>Fide Master
>Lover of Chess

I will believe that comps are GM strength when I see a comp win a match against
one.

James



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.