Author: Vincent Diepeveen
Date: 12:47:35 12/26/02
Go up one level in this thread
On December 26, 2002 at 14:58:51, Ed Schröder wrote: >Merry Christmas to all CCC members in the first place! > >As you know I have retired from chess programming, after 22 years enough is >enough. Right now I am documenting my work, part of it will be a description of >the main parts of REBEL in the hope some others will benefit from it. > >The main page can be found on: > >http://members.home.nl/matador/chess.htm > >The programmer stuff on: > >http://members.home.nl/matador/chess840.htm Thanks for your excellent contribution to computerchess Ed. It's a very great thing to do. You describe a mate killer move giving a 3-4% speedup. I was amazed by this statement. Around 1996 Bart Weststrate described this mate killermove to me too. I then tried it in DIEP. Indeed it speeded DIEP up bigtime at testsets. Some positions it searched plies deeper in fact. However after a few weeks i saw diep struggling to get another ply in games. In game positions the where there was not a mate trick to find and some stupid line went into a mate once, then being saved to the mate killer table; (and usually you do not get into a position where the opponent is getting mated soon) there DIEP suffered *bigtime* from this special killertable i had. So i threw it out. Note that i indexed it a bit different than you do, but that won't change the picture at all. The 3-4% speedup you claim here, is that based upon tactical testsets? Best regards, Vincent >So far only move-ordering is finished but I am planning to have the rest of the >issues ready at least end of next month. > >My best, > >Ed
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.