Author: Vincent Diepeveen
Date: 08:35:33 10/17/02
Go up one level in this thread
On October 17, 2002 at 05:17:50, Gian-Carlo Pascutto wrote: if white takes in a slightly different position than this, it is a win for black. If white doesn't take, most likely white easily draws this. Engines should not take unless there is clear proof that a position is won. Just 2 far passers in the evaluation doesn't make sense at all. I have seen too many pro games where exchanging was just won for black in similar positions. Suppose you can eliminate 1 passer far in the horizon. Look it is like 15 moves to get a clear won position here with white. However the human insight can detect that very easily here, but if you can eliminate a passer at move 14 and still be in time at the other passer, then you have a major problem and lose after rxf3. Obviously it is for the far future generation of programs to find such key moves. If they do, they'll be very hard to beat for grandmasters. Because if they find such moves, it's not easy to take into account that in a few years of time we search plies deeper too. So you see it plies in advance. DIEP doesn't take here by the way. Perhaps some future version will. It's not trivial knowledge to make though. Otherwise it would be in DIEP already. >On October 17, 2002 at 04:51:28, Christophe Drieu wrote: > >>[D] 8/4k3/4p3/pp5p/6p1/2P2rP1/PPK2R1P/8 w - - 0 1 >> >>Rxf3 ! > >Good question. This is connected central passers vs two >outside passers. The outside passers wins almost always. > >You can see clearly here which programs have this knowledge >and which ones not. I'm amazed how many pros don't. > >I'm also amazed mine won't play it. It has the knowledge. >A bug or is there more than meets the eye? > >-- >GCP
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.