Article View: gnu.chess
Article #5847Re: What IS the strenght of GNUchess
From: Richard Title
Date: Tue, 12 Nov 1991 14:37
Date: Tue, 12 Nov 1991 14:37
28 lines
1832 bytes
1832 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
A while back some of us did some strength testing of gnuchess, because we were
considering it as a base for a chess project...
The documentation that comes with gnuchess claims a rating of around
2300 on a Sparc, but this is a lie.
In our tests, the Fidelity Mach III beat gnuchess decisively, by a score of
6-0. The older Novag Superconstellation was slightly better
than gnuchess (2.5-1.5), and the Mephisto Mondial 68000 defeated it 1-0.
The above experiments were conducted on a Sparcstation, at a time
control of 1 minute/move, with gnuchess version 3.1+.
We were planning on more extensive testing, but gnuchess was so weak
we didn't feel it even worth the time to go on. It appeared to be
about 1800 strength. I'd be happy to post actual games if people desire.
Rich
Technical Details
Message-ID:
<khvpq3INN62f@early-bird.think.com>Path:
arc1-ca.pugleaf.net!gmdzi!zeus.ieee.org!sdd.hp.com!think.com!titleReferences:
<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 06:50:07 PST Path: arc1-ca.pugleaf.net!gmdzi!zeus.ieee.org!sdd.hp.com!think.com!title From: title@think.com (Richard Title) Newsgroups: gnu.chess Subject: Re: What IS the strenght of GNUchess Message-ID: <khvpq3INN62f@early-bird.think.com> Date: 12 Nov 91 14:37:55 GMT References: <1991Nov12.085731.784@sics.se> Organization: Thinking Machines Corporation, Cambridge MA, USA Lines: 28 NNTP-Posting-Host: hespera.think.com