Author: enrico carrisco
Date: 18:13:07 12/28/05
Go up one level in this thread
On December 28, 2005 at 13:58:28, Jonas Cohonas wrote: >On December 28, 2005 at 03:16:39, enrico carrisco wrote: > >>http://www.computerschach.de/index.php?option=com_content&task=view&id=208&Itemid=167 >> >>-elc. > >" To obtain relatively demanding games, I chose the time control with 10 minutes >per game and 10 seconds per move." > >I must admit that time control worries me when the tester also states: > >"I wanted to create a series of tests which make it possible to eliminate the >weak points mentioned above and to make a really comparable Rating List. >Furthermore it was my aim that this Rating List with relatively little effort >can be updated at the inclusion of a new engine. To avoid problems at the >automatic engine tournaments, engines, which work without problems on the >Chessbase-GUI, are exclusively tested" > >Sounds to me like this list will give you a good estimate, but if you want more >accuracy and more thurough rating list, then either the CEGT or the SSDF. I disagree. I value the information from all three of those sources, but I believe the CSS Rating List is the most systematic. Consider the following: * Same set if openings in all matches. * Same set of opponents (each new engine faces.) CEGT tests have much nore variance because the list of opponents is not identical (some programs are played against more than others), multiple openings used (varied by tester), speed differences in machine speeds (though mostly accounted for in adjusting the time control of the match), multi-threaded engines included and multiple versions of one program (including different settings of the same versions.) However, this, in my opinion, also makes CEGT very interesting and I value its results. However, I would not assess those factors to provide "more accuracy and thoroughness" as you state in your posting. Regards, -elc.
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.