Author: Pete Galati
Date: 18:02:03 06/06/01
Go up one level in this thread
On June 06, 2001 at 20:36:50, Dann Corbit wrote: >On June 06, 2001 at 19:50:31, Mark Schonfeld wrote: > >>Does anyone know of any larger test suites then the WAC set that is available? >>Can anyone explain to me why over time these test suites have not been added to? >>It appears with so many analyses on various chess positions being done in here >>that maybe some of the results should be aded to the test suite or is this being >>done? Thank you in advance for your help. > >Lots of big ones are available. WAC is not a particularly large one. The value >in test suites is not in the size, but (far more so) in correctness. > >I think WAC is what people should use in beginning to write a chess program. >Eventually, they should move up to LCT II. I don't think you can judge how >strong a chess program is from test suites. They are a necessary (but not a >sufficient) condition for a chess engine to be powerful. > >There are no other test suites that are debugged as well as these, I think. MES >(in particular) is full of gaffes. > >Have a look here: >ftp://cap.connx.com/pub/EPD/ > >And here are 20 million analyzed EPD records: >ftp://cap.connx.com/pub/Public_CAP_Results/mega/m2000.xtd.bz2 > >Its 500+ megabytes compressed and 3.5+ gigabytes decompressed. That's only 10% >of the actual EPD data, but it does not make much sense to publish the whole >pile for obvious reasons. Do you have favorites? How about anyone else? Pete
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.