Author: Laurence Chen
Date: 16:49:16 12/31/00
Go up one level in this thread
On December 31, 2000 at 16:26:23, Andy Serpa wrote: >It seems most programs are getting better and better at avoiding blocked >positions. Does that mean that are getting worse and worse at analyzing blocked >positions that are presented to them? > >In other words, if I want to use a computer to help analyze my own human - human >games in which I might get into very closed positions on purpose, and where >trying my best to unblock it might not be the best thing to do, is a computer >any use to me here? Is it just going to come up with unblocking moves no matter >what, or is there any program that can do a decent analysis of such positions? >(Positions that the computer would never find itself in had it played the game) > > >Andy Serpa >chessmad@chessopolis.com >http://www.chessopolis.com A blocked position must become open sometime, it cannot remain closed forever. If one create a Stonewall, and there's no possibility of opening up the position then the game will end up in a draw. Pawn breaks are a necessity in blocked positions. I think that blocked positions are mostly common in the Class B or Class C games between human players. The reason could be that at this level the players don't know when to execute a pawn break and misses many opportunities to do it. The chess engines approach to aim for opening up the position is correct, however, the problem is that very few chess engines are capable of maneuvering the pieces to the correct squares to open up the position. IMO, Gambit Tiger would be the best engine for analysis in closed positions, this is based on my observation of games played by this engine against other chess engines. This may not necessary be true against a human opponent. Against human opponents, Deep Fritz seems to have knowledge of how to play closed positions, this is based on the games which Deep Fritz played against other human GMs. Then again closed positions don't offer many opportunities to score a full point. Regards, Laurence
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.