Author: scott farrell
Date: 17:13:36 12/07/02
Go up one level in this thread
On December 07, 2002 at 12:04:46, Colin Frayn wrote: >I've fixed the problem with recognising trapped rooks... sort of. Beowulf now >recognises the following positions (and others like them) as completely trapped >rooks (and penalises them based on game stage); OK, sounds good. How did you do it? what was your theory to detect it? Do you have any pseudo code for those interested? Scott > >[D]5k2/p2p3p/2p2pp1/3p3r/1P1N1P1P/6P1/P5B1/4K3 b - - >Rh5 > >[D]r4k2/2R2p1r/4pBp1/4P2p/pR3P1P/Pb2K1P1/8/8 b - - >Rh7 > >[D]8/7p/6k1/5p2/p1p2P1R/P2pPK1P/1r6/8 w - - >Rh4 > >Beowulf even reognises more complicated positions than these, such as ones with >nasty exchanges, but not ones where a complicated series of pawn blocks prevents >the rook from escaping, but the rook has a significant freedom of movement. > >I added in 'trapped knight' tests too, as they're also fairly easy to do. > >Of course, it's a bit slow now ;) >I don't really care though as I'm now going for the 'slow, but complicated' >approach. I figure that if you can spot tactics two plies earlier then it >doesn't matter if your eval() is an order of magnitude slower. > >Cheers, >Col
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.