Author: Leen Ammeraal
Date: 10:40:57 08/30/02
Go up one level in this thread
On August 30, 2002 at 10:39:15, Will Singleton wrote: >On August 30, 2002 at 07:54:24, Leen Ammeraal wrote: > >>After spending a whole morning looking for an >>error in my program in connection with the >>position CMB6 in the LCT II testsuite, I >>finally found the problem. However, it was >>not in my program but in the line for that >>position: >> >>2r1k2r/2pn1pp1/1p3n1p/p3PP2/4q2B/P1P5/2Q1N1PP/R4RK1 w q - bm exf6; id >>"LCTII.CMB.06"; c0 "Lilienthal - Capablanca, Hastings 1934"; >> >>As the beginning of this line shows, there is a >>black rook on c8. In spite of this, the q in >> >>w q - bm >> >>indicates that black still has the queen side >>castling right! In view of this, this q should >>be replaced with a hyphen (-) or a k. >> >>If you ever use this LCT II testsuite, you should >>correct this line. >> >> >>Leen > >Thanks for pointing this out. I checked my files, and some copies had the k, >some had q. > >But I'm wondering why it would make any difference to your program. First, >Queen should realize that the rook is out of position and cancel the castling >rights, or if the rook moved back to a8 that would cancel also. But more >importantly, it appears to me that castling doesn't come into play in the >solution. > >Maybe you still have a bug? > >Will Well, my program trust the correctness of the epd line (which may be unwise). Then it tries queenside castling since it considers it as a valid move simply on the basis of the castling right flags, making a mess of its internal datastructures. In a normal game the castling right flags are always correct. After loading a position the program just takes the castling rights from the input line. Leen
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.