Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: I will hazard a wild guess that a computer can be a GM at G/60

Author: Robert Hyatt

Date: 20:32:14 02/17/00

Go up one level in this thread


On February 17, 2000 at 18:33:20, Dann Corbit wrote:

>The morbid fear of playing on against Deep Junior and the "true" result against
>super GM players shows that (I think) at G/60, a computer does play like a GM.
>Maybe even a super


3-4 years ago, I would have said "You are as full of crap as a Christmas
turkey..."  But since then, I have greatly revised my opinion here. It
started when Crafty, Ferret, ChessMaster, and one other program played in
a round robin event on chess.net with 4 GM players.  The time control was
not sudden death I don't think, but was 30 something where something was
very short (it _might_ have been 30 0, but I don't remember).  In that
event, every computer played every GM.  And at the end, all the computers
had better scores than the best GM.  IE we owned the top 1/2 of the places,
the GMs owned the last 4 places.

After a few more such events that I personally played in, I became convinced
that at 30 0 and 60 0, GM players have _great_ difficulty.  They do one of
two things:  (a) the press too hard early to try and win before time becomes
an issue, and this usually fails;  (b) they get low on time and then get badly
out-blitzed by the computer.  I don't think computers can beat _ALL_ GM players
at that time control, because I know a couple that are _serious_ problems since
they understand computers very well.  But most are going to roll over at 30 0
and 60 0.

At longer time controls, particularly non-sudden-death time controls, time isn't
as much of an issue, and the game doesn't resolve into a blitz match which a GM
most likely can't win.

We are probably seeing the "end" of the welcome-wagon for computers in human
chess, particularly at faster time controls.







>
>I think that Amir and Shay should be justifiably proud of their effort.  Any
>machine+program that can strike fear into the heart of a super-GM is an
>astonishing thing.
>
>I think that Adams is not to blame for any controversy.  If I were a player, I
>would always push for any edge I could get, even with the arbitration committee.

I have chatted with him a few times on ICC.  He didn't seem like the type to
behave as he appeared to behave.  I believe that KC was the _real_ problem here,
with no idea of how to manage such a tournament, no idea of how to anticipate
the problems and have written protocols ready to handle them, rather than resort
to a poorly-conceived knee-jerk reaction that made the entire KC enterprise look
like a bunch of amateurs.

I begin to suspect Adams was just in the wrong place at the wrong time.  It
could have been _any_ GM on the other end and KC would have screwed this up just
as badly, unless both ends were GM players.






>
>Any shame (and there is clearly some to go around) goes to the arbitration
>party.
>{purely}IMO-YMMV.



At least to the person(s) responsible for a decision that was nothing short
of ridiculous.




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.