Author: Ron Murawski
Date: 20:58:27 07/28/01
Go up one level in this thread
On July 28, 2001 at 22:22:23, Bruce Moreland wrote: >On July 28, 2001 at 18:47:21, Ron Murawski wrote: > >> >>I have implemented a pin bitboard for king-pinned >>pieces and it has helped the strength of my program. > >I don't know anything about the rest of this, because I have never identified >pinned pieces. How do you know that it helped the strength of your program? > >bruce > My program kept losing games by overlooking the fact that pinned pieces cannot capture. A good-seeming move was sometimes not so good. For situations where the search didn't reveal the lost power of the pinned piece the "best" pv was definitely NOT the best. I've got a slow-searcher and routinely get out- searched by up to 4 plies. Keeping track of the pinned pieces was as good as gaining several ply of look-ahead and "discovering" a pv gone bad. I had the most problems in games where the program failed to castle and got killed by pins early in the game. The program now knows to avoid pinned pieces whenever possible before damage can be inflicted by the opponent. The program now looks to avoid pins through king moves or by the interposition of pieces that "fight back". The program is new enough to have no rating. I've just observed its behavior and made adjustments that seemed appropriate. I can't prove that the pin bitboard increased its strength, but the program's defensive skills seem better and it has become more difficult for opposing programs to checkmate it. In other words, it still loses, but now it takes more moves to do so. I'm new to chess engines, so take what I say with a grain of salt. I'm changing so many things on a daily basis that it's difficlt to determine exactly what change has caused what increase/decrease in strength. I'm also using non-traditional methods in the eval, so what is appropriate for me might not translate into something that makes sense with what you are doing. Ron >>My question is: Is it worthwhile to identify ALL >>pinned pieces? >> >>Whereas the king-pinned pieces were quite easy to >>determine and very important to the scoring, any >>queen-pinned, etc. pieces will take much longer to >>calculate and have less effect. In fact, some of the >>pins seem to be phantoms as they might disappear >>because of possible checking moves, tactical threats, >>etc. >> >>Is it better to discover these pinned pieces in the >>search, or is it better to do all the additional >>processing in the evaluator? Has anyone tried >>implementing the detection of all pinned pieces and >>was the time spent looking for these pieces worth the >>effort? >> >>Thanks in advance, >>Ron Murawski
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.