Author: Miguel A. Ballicora
Date: 15:32:34 03/24/02
Go up one level in this thread
On March 24, 2002 at 15:55:25, Uri Blass wrote:
>On March 24, 2002 at 15:07:09, Peter Berger wrote:
>
>>The time for the "big" match has come - the first game will be played today :
>>Sunday, 24/03/2002, 4 PM EST at ICC (=server time) ( http://www.chessclub.com).
>>
>>The contenders:
>>
>>Gaviota 0.30, the young and rising amateur chessprogram star from Argentina
>>(more information and download at http://www.msu.edu/~ballicor/gav/)
>>on a fast AMD Athlon 1.33 GHz
>>
>>against its very programmer:
>>
>>FM Miguel A. Ballicora
>>
>>Gaviota will play with handle gaviota(C) ( thanks to Andrew Mc Farland aka
>>"Zek", ICC admin). Game number will be announced in channel 64 when the game
>>starts ("+ch 64").
>>To observe the game type "obs gaviota" or "follow gaviota" to be sent to the
>>board once the game has started - you can connect and watch as a guest if you
>>don't have a handle.
>>
>>Program and human will battle it out in a 4 game match with 30/5 time control (
>>game in 30 minutes with 5 secs increment/move).
>>
>>There will be one game today - the next rounds will be announced soon.
>>
>>A tough match is to be expected as the opponents might be of quite similar
>>strength ?!
>
>
>I think that gaviota is a better player at the time control that they are going
>to play but Miguel has the advantage of knowing the opponent so I am not sure
>about the result.
Do I have to feel flattered as a programmer or annoyed as a player? :-)
Just kidding. Thanks to the all the guys who wished me luck.
Well, Gaviota just crushed me. I put my head right into the lion's mouth and
gaviota played flawlessly.
I did a terrible election of move sequence allowing the Volga/Benko Gambit.
When I was in it, I had to go for variations that I never played to avoid
the mainstream variations of the volga that are excellent for computers and
I do not feel comfortable in those situations. Good job Peter that must have
prepared a good book and good job Gaviota. I have to be more careful.
About knowing my opponent, it won't be a big problem because I won't go for
weird tactics or "antinatural" anticomputer tactics. Besides, I do not control
the book. I will, of course, be careful with tactics but that is very general.
I did a lousy job at that today.
>>FIDE lists Miguel with Ballicora, Miguel f ARG 2245, but he says he hasn't
>>played seriously for a longer time.
>>
>>Gaviota has participated in a few computer tournaments.
>>
>>Andreas Herrmann lists Gaviota with 50.( 49.) Gaviota 0.3.0 2230 N 219 109 26 84
>>122.0 55.7 2189 N in his rating list.
>
>I do not trust rating lists for computers and an experiment in the winboard
>forum showed that programs simply get worse rating than the rating they deserve.
>
>I think that you should add 100 elo to the rating of programs in Andreas
>Herrmann list to get a better estimate for their real rating at 120/40 against
>humans who are not prepared to take advantage of the weaknesses of the spesific
>programs.
Hard to say. You might be right. In the winboard forum there were very few games
to make a conclusion now. I do not agree with you (at 120/40) but
it is just a guess from my side.
Regards,
Miguel
>
>Uri
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.