Author: Uri Blass
Date: 12:34:20 09/09/01
Go up one level in this thread
On September 09, 2001 at 14:36:07, Gian-Carlo Pascutto wrote: >On September 09, 2001 at 13:54:03, Uri Blass wrote: > >>The problem is that I cannot prove even +1.xx and even in the last positions in >>the line that I got by email I can see only 0.xx evaluation(I do not care about >>the fact that black has material advantage of 1 pawn in the end of the line >>because programs can see that white has some positional compensation for the >>pawn). > >As I said, there is no guarantee at all that I had the best moves >for black in the line I emailed. I just picked something where black >wins. > >So you don't agree black wins there. Well what a surprise. > >>Singular etensions are extensions for things that you can prove. >>If in almost every node there are 2-3 logical lines then it is clear that >>singular extensions can prove nothing. > >My implementation of SE for crafty (which is suboptimal) finds >that black is ahead (it doesnt see the win of the piece -yet-) >three plies earlier than normal crafty. ahead is not winning and 3 plies are not 20 plies. 20 plies difference does not make sense. deep Fritz also agrees that black is ahead after Bg5 and it believes that white did a positional mistakes some mobves before. > >As I said before. It's not because _you_ can't see something that >it isn't there. It is more logical to assume that it is not than that it is. It does not make sense to assume that Deep thought could really see winning 2 pawns. Unfortunately I have not the email of hsu (otherwise I would ask him directly about it). Maybe we can solve the problem if hsu says that the +2 was result of a bug or of a big positional score than we can solve this problem. Uri
This page took 0.01 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.