Author: Marc Bourzutschky
Date: 12:30:42 02/02/05
Go up one level in this thread
On February 02, 2005 at 13:19:11, Anthony Cozzie wrote: >On February 02, 2005 at 13:14:21, Dieter Buerssner wrote: > >>On February 02, 2005 at 13:00:51, Anthony Cozzie wrote: >> >>>On February 02, 2005 at 01:06:30, Robert Hyatt wrote: >>> >>>>>KRPKRP (only the white-side is available from Bob right now?) >>>> >>>>think about it. If pieces are symmetric you only need one side. Just turn your >>>>white pieces into black ones, and the black ones into white ones, and look it >>>>up. :) >>> >>>doh :( >> >>Bob forgot to mention the mirroring needed for positions with pawns. The pawn >>TBs are not identical (they have the identical contents, of course, ordered a >>bit differently). >> >>C:\tb>dir kpkp* >> >>05.08.2003 12:00 1.111.988 kpkp.nbb.emd >>05.08.2003 12:00 1.116.299 kpkp.nbw.emd >> >>I think, current Nalimov code needs both (which may be not intended). From the >>code: >> >> TB (kpkp, false, false, false, (T22<P, P>::IndCalcW), (T22<P, P>::IndCalcB), >>3863492, 3863492) >> >>The "fsym" is set to false. >> For the 33p endings the redundancy is eliminated, with IndCalcB replaced by IndCalcBF which does the appropriate flipping. That is something I added to the code. Eugene was skeptical at first that this would work, but he seems to have come around :-) For kpkp the redundancy is no big deal, so there was no need to eliminate it. -Marc
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.