Author: Robert Hyatt
Date: 07:39:22 11/01/02
Go up one level in this thread
On November 01, 2002 at 07:22:17, Peter Berger wrote: >On October 31, 2002 at 11:22:33, Robert Hyatt wrote: > >>Hard to answer, but I can't imagine that KRR vs KRN is winnable very often, >>based on the >>.tbs file on my web site. Almost all are draws. Of course, without the KRRKRN >>egtb, >>a wrong move can result in a loss, but in gereral it is drawn, almost all the >>time. >> > >I think the problem is the 0.00 evaluation here that seems to make all moves >equally attractive for Crafty while at the same time the knowledge of the >opponent to drive the opponent king to a corner is sufficient to make life >difficult for Crafty. In fact the opponent didn't even try very hard to force >Craftys king to the corner - it just went there. >I expect the numbers of lost positions to be higher when the defending king has >been driven to the final rank or a corner of the chessboard. > >Of course with 6-men tablebases this is no problem but I dont expect them to >become very common soon as they are so big. > >Thanks for the tablebase information posted. I'll look at that. It is supposed to say "krr can't win vs krn" but not set the score to absolute zero. Otherwise, you are obviously correct, in that this will lead to the weaker side getting pushed around... I'll look to see if the mate evaluation is somehow getting squashed in this case, which it must not...
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.