Author: Mark R. Anderson
Date: 03:53:56 02/11/05
Go up one level in this thread
On February 11, 2005 at 02:31:23, Kurt Utzinger wrote: >On February 10, 2005 at 21:30:40, George C Williams wrote: > >>Or solves much faster then other programs. > > Trying to prove something with a test position > is not the right way in my opinion. What counts > are the games played vs a lot of other engines > and the overall score of a chess program. And > in this respect I am almost sure: Shredder 9 > will be top. > Kurt Kurt, I agree with you. It's always possible to find a position that any one of our engines finds quicker than the others. What matters is how the engine *plays* in practice! For instance, using the "Marathon" set of test positions (210 positions from CME), I tested most of my engines, and got the following unexpected result: #1 Tao 5.6 (!!) #2 Gothmog 1.0 #3 Ruffian 2.1 #4 Pro Deo 1.0 (haven't tested 1.1 yet) #5 Hiarcs 9 ...etc Shredder 9 was well down in this pack, something like 13th place, but it kicks all of the other engine's butts in actual play. Also, Tao is a good engine, but can't begin to compete with the best of the commercial engines or even the best free engines (list List 5.12, Pro Deo 1.1, Aristarch 4.5, etc). So, positions are nice, but don't really tell the story about the strength of an engine. So far, I am checking Shredder 9 UCI against Junior 9, Hiarcs 9, Fritz 8, Gandalf 6, Ruffian 2.1, List 5.12, Pro Deo 1.1, Tiger 15, Deep Sjeng 1.6, etc, and Shredder 9 is kicking everyone's butt. Good work Stefan! Yes, early results only, but they are really excellent. I may not be able to put an exact number on the elo increase yet, but it's very obvious that it's a true and good strength increase to an engine that was already the strongest. Mark
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.