Author: Uri Blass
Date: 03:38:25 06/08/03
Go up one level in this thread
On June 08, 2003 at 06:00:39, Manfred Meiler wrote: >Hello, > >at http://www.computerschach.de/test (link "Weltmeister-Test") there's an >updated version of my EXCEL sheet (german and english version) with the results >of 187 engines in the test suite "Weltmeister-Test" (WM-Test) - hardware: AMD >Athlon Thunderbird 1400 mhz. > >This test suite was developed by Dr. Michael Gurevich und Heinz-Josef Schumacher >and includes 100 test positions from games of the different (human) chess world >champions: 38 positions in King attack, 36 positions in positional playing and >26 positions in endgame. >Each of the 100 test positions was tested 20 minutes in analyse mode, without >access to opening books, but with access to endgame databases (Nalimov >tablebases). > >Added is a collection of the 100 test positions in the formats cbh, cbf, pgn and >epd. To all test positions there are variations and comments (in german) for >better understanding the goals of the single position. > >Since my last published EXCEL sheet (2002 December, with 140 tested engines) the >following engines were tested by me: > >- Fritz 8,0,0,8, Fritz 8,0,0,23 >- Shredder 7.04, Shredder 7 SE, Shredder 7.04 Gambit, Shredder 7 El Cid, Triple >Brain "Shredder 7/Ruffian 1.01" and "Shredder 7/CM9_Pillen" >- Junior 8, Deep Junior 8, Deep Junior 8.ZX, Deep Junior 8.YY >- CM9000_SKR, CM9000_Grailmaster7 >- Ruffian 1.0.5, Ruffian 2003-04-13, Ruffian 2003-05-06 >- SOS.3 for Arena >- Deep Sjeng 1.0, Deep Sjeng 1.5 >- Aristarch 4.21 >- Yace Paderborn >- Anaconda 1.0 >- Little Goliath Nemesis 1.006 >- Zarkov 4.62 >- LambChop 10.99 >- Delfi 4.1 >- Green Light Chess 3.00 >- Movei 0.07.99 >etc. > >Manfred Some comments: 1)I am not surprised that movei is one of the worse programs in king safety. The public version that you tested has no evaluation of king safety(latest version has some king safety evaluation). 2)I am surprised by the relative good result in endgames(not last place and even better than delfi4.1) Movei does not use hash tables for pruning and the public version also does not know about passed pawns. I suspect that the problem may be that programs can find the right move for the wrong reasons because I do not believe that movei is even close to delfi in endgames. 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.