Article View: gnu.chess
Article #5848Re: What IS the strenght of GNUchess
From: Roy Brunjes {Systems Eng} Sun Ro...
Date: Tue, 12 Nov 1991 18:57
Date: Tue, 12 Nov 1991 18:57
47 lines
3065 bytes
3065 bytes
In article <1991Nov12.085731.784@sics.se> roland@sics.se (Roland Karlsson) writes:
>
>Some say that GNUchess is a very strong chess program. I think I
>have heard that it should be about 2500 US ratings on a SPARC station.
>Some others say that it plays an awful chess. Far below commercial
>chess computers. Those two statements are contradictory. What is
>the truth? Has anyone made (serious) testings of the strength?
>
>
>Roland Karlsson
>--
>Roland Karlsson
>SICS, PO Box 1263, S-164 28 KISTA, SWEDEN Internet: roland@sics.se
>Tel: +46 8 752 15 40 Ttx: 812 61 54 SICS S Fax: +46 8 751 72 30
I conducted a 20 game match between GNUchess 3.1 (not 3.1+, sorry) on
a SPARCstation 2 and my dedicated chess computer (at the time) the Novag
Super Expert C. The results: A victory for the Novag by the score of
10.5 to 9.5. That would put the rating of GNUchess somewhere around
high 2100's USCF (about low 2000's FIDE?? I'm no expert on such conversions).
Running GNUchess 3.1 on a SPARCstation 1, the Novag won convincingly, though
I can't recall the exact score of the 20 game match there.
Oh, the time control for both matches (GNU on SS2 vs Novag Super Expert C, and
GNU on SS1 vs Novag Super Expert C) was Game in 1 hour. I set GNU to play
60 moves in 60 minutes and considered that to be game in 1 hour, though in
actuality, sometimes the games went on past 60 moves, and then GNU got more
time while the Novag (having a true Game/n moves time limit) still had to live
by the time constraint of game/60.
I feel confident in saying that GNUchess is nowhere near 2300 strength on
these platforms. In current testing between the Mephisto Lyon 32 bit and
GNUchess on a SPARCstation 1+ (SPARCstation 2 not available at the moment),
I set GNUchess 3.1 to 60 moves in 60 minutes and the Lyon at Game/30
(not a typo! Game/30) and thus far, Lyon has won 2 and drawn 2. Too early
to draw many conclusions though. I suspect that Lyon at Game/30 will prove
stronger than GNUchess at 60/60; just a guess at this point.
Hope this helps you determine an approximate rating range for GNUchess.
--
Disclaimer: Any opinions expressed are probably my own, but certainly NOT
those of my employer.
Roy Brunjes roy.brunjes@East.Sun.COM
Technical Details
Message-ID:
<569@playroom.UUCP>Path:
arc1-ca.pugleaf.net!gmdzi!zeus.ieee.org!sdd.hp.com!cs.utexas.edu!sun-barr!cronkite.Central.Sun.COM!newstop!eastapps!playroom!gambit.East.Sun.COM!royReferences:
<1991Nov12.085731.784@sics.se>Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 10d71d,81a388b9f8acdd17
X-Google-Attributes: gid10d71d,public
X-Google-ArrivalTime: 1991-11-12 11:22:17 PST
Path: arc1-ca.pugleaf.net!gmdzi!zeus.ieee.org!sdd.hp.com!cs.utexas.edu!sun-barr!cronkite.Central.Sun.COM!newstop!eastapps!playroom!gambit.East.Sun.COM!roy
From: roy@gambit.East.Sun.COM (Roy Brunjes {Systems Eng} Sun Rochester)
Newsgroups: gnu.chess
Subject: Re: What IS the strenght of GNUchess
Summary: GNUchess 3.1 on SPARCstation 2 much less than 2300
Keywords: GNUchess strength overrated
Message-ID: <569@playroom.UUCP>
Date: 12 Nov 91 18:57:42 GMT
References: <1991Nov12.085731.784@sics.se>
Sender: news@playroom.UUCP
Organization: Sun Microsystems, Inc. Rochester, NY
Lines: 47