Author: GuyHaworth
Date: 00:03:51 09/23/04
I see that Eugene has flagged the fact that he is dealing with some errors in the .tbs files. These have been spotted by: a) comparing DTM EGT stats with DTC, DTZ and DTZ50 stats, b) comparing stats with the index-ranges in the generator code Years ago, an error in the KBPKN stats appeared. The Hyatt .tbs file is still wrong. This error is a 'one off' as far as I know. More systematically, with the big numbers found in 6-man positions, 2^32 or multiples of 2^32 positions get dropped when there is a big count. Obviously, a consequence of using 32-bit integers. For 6-man pawnless endgames, only KBNKBN is affected: add 2^32 to the draws. For 4-2p endgames, some 28 of 95 endgames are affected. In an extreme case, the KRBPKQ .tbs does not count in 3*2^32 positions. For 3-3p endgames, some 7 or the 17 endgames to date are affected. This is not going to stop the world for too many people, including me, and Eugene is dealing with it. Guy
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.