Author: Uri Blass
Date: 01:05:55 10/21/03
Go up one level in this thread
On October 21, 2003 at 01:12:02, Stephen A. Boak wrote: >On October 20, 2003 at 16:47:45, Uri Blass wrote: > >[..] > >>I do not claim that I am always right but the fact that GM's do not find >>something is not a proof and it may be possible that a good practical decision >>is not to investigate it. >> >>It may be a bad idea to work months only to try to find if there is a saving >>line that you do not know if it exists when you can at the same time improve >>your knowledge in other openings. >> >>GM's are not scientists and they try to achieve the best results that they can >>achieve in games and not to be sure about the theoretical result of some >>position. >> >>Uri > >If I lose my queen (White or Black pieces, doesn't matter) by a blunder on move >3 or 4, all GMs would probably say I'm lost. I would, despite the inability to >*prove* it by exhaustive searching of *all* variations. > >But until you [Uri] calculate *all* subsequent variations, would not say I am >lost for sure? Only that I am probably lost? :) > >Because chess is not *solved* for most positions (except for rare positions with >forced mates or forced repetitions, etc, and endgames through 5 pieces, some 6 >pieces), it is not possible to be sure of anything (by exhaustive research on >exponential quantities of variations). You can't be sure that a position is >lost, even where a queen is blundered after 3 or 4 moves. > >If the above paragraph is true, why state the obvious just to disagree with >someone whose opinion (and perhaps research) concludes the opposite. > >--Steve I did not say that you have to see a forced checkmate to say that you lose the game. The question is if there is a reasonable doubt about the question. If my opponent has queen advantage or even knight advantage with no compensation in the middle game there is no doubt of even 0.1% that the theoretical result is a win for the better side. In the case of the relevant game there is a serious doubt because even proving +2 with Fritz8 is very hard and you need to analyze thousands of lines for that. Of course Fritz8 is not god and there are positions when I can be sure of winning inspite of evaluation of the computer based on knowledge but there are mainly endgames when there is a long plan and not complicated middle game position when I always suspect that there may be something that I miss. The main problem here is that computers still do not know to do statistical analysis and say line a is losing,line b is losing,line c is losing so black is probably losing after b4 even without finding a forced line. Uri
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.