Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Swiss Test - Ranking

Author: Uri Blass

Date: 09:02:26 11/09/02

Go up one level in this thread


On November 09, 2002 at 11:40:08, Mike S. wrote:

>On November 09, 2002 at 10:33:10, Uri Blass wrote:
>
>>On November 09, 2002 at 08:15:24, Walter Eigenmann wrote:
>>
>>>Hi
>>>
>>>The new SWISS TEST contains 40 "avoid moves".
>>>
>>>Ranking list so far (Athlon XP2000):
>>>
>>>01. Gambit Tiger 2.0......30
>>>02. Junior 7..............29
>>>03. Shredder 6.02.........25
>>>04. Fritz 7...............24/4
>>>05. Hiarcs 8..............24/2
>>>06. Nimzo 8...............23/3
>>>07. CM8000/Pillen.........23/2/11
>>>08. Pepito 1.55...........23/2/10
>>>09. Ruffian 1.0.1.........23/2/8
>>>10. Dragon 4.3.2..........22/3/8
>>>10. KingOfKings 2.02......22/3/8
>>>10. Madeleine 0.1.........22/3/8
>>>13. LG2000 3.8............21/3
>>>14. Crafty 19.01..........21/2
>>>15. Terra 2.9.............20/4
>>>16. Mint 2.3..............20/3
>>>17. Comet B50.............20/1
>>
>>
>>This shows the value of the test.
>>
>>Mint is not going to be better than Comet
>>in every logical test.
>
>As usual, your comment is sceptic. - But don't forget, a test tells factual
>results from a specific set of positions. It just means, in that very test Mint
>solved as good as Comet (avoided wrong moves).

A normal chess program at Mint's level
is not going to get a better score than a normal chess program
at Comet's level in test suites like the GCP test suite.

This test suite is a joke because programs can often solve
it for the wrong reason.

The problem is not the fact that the
test is to avoid moves but the fact
that programs can often avoid the move for the wrong reason.

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.