Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Hunting Hydra

Author: Drexel,Michael

Date: 08:09:58 06/27/05

Go up one level in this thread


On June 27, 2005 at 09:01:45, Peter Skinner wrote:

>On June 27, 2005 at 03:10:02, Drexel,Michael wrote:
>
>>I don´t understand your argumentation:
>>
>>Shredder is tuned for computer (absolutely correct)
>>
>>Hydra is tuned for human players (I don´t fully agree, it was probably only
>>tuned for the Adams match)
>>
>>I don´t understand your conclusion Shredder therefore would get killed in such >a match. AFAIK know Hydra is a computer.
>>
>>Are you better informed than we are. Is Hydra a Turk with Bobby Fischer inside?
>>:)
>>
>>Michael
>
>Shredder has already been destroyed once by Hydra.
>
>Sure the excuse has been offered that Shredder was a beta version and was
>extremely buggy. (Why play such a match with a non-stable version escapes logic,
>but ok... I'll buy it).
>
>Even taking into account the "buggy version", Hydra dominated Shredder. It
>simply played at a level that a non-buggy Shredder could perform at. It is
>better tactically _and_ positionally.

It was positionally stupid as hell. Look at the Topalov-Hydra game where Hydra
was better after the opening and got outplayed.
The 1...a6 game posted by Pablo above (maybe only played with 1 FPGA card)
also isn´t very impressive...

Maybe you only think it was better positionally. In reality Shredder was
inevitable lost in 2 games after it had left the book.

>
>I would be willing to wager my kids to a Martha Stewart child labour factory
>that even a "stable" Shredder wouldn't take a single game from the same machine
>that faced Adams.

I assume this machine has already lost to Shredder during their tests.


From games that I have seen Shredder play at 120 0 on ICC vs
>GM's way below the level of Adams, Hydra would kill it positionally.

Shredder has a very optimistic evaluation which seem to help against other
engines but is sometimes a drawback against humans.

Shredder certainly would have to be optimized to run on 8-way Opterons at least
in order to have decent chances.

Michael

>
>While I would love to see the match happen, I think Junior would be a much more
>challenging opponent for Hydra. It is a positional program, and isn't just tuned
>for computer play.
>
>Peter



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.