Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Rebel - IsiChess: Some notes

Author: Uri Blass

Date: 12:57:54 10/20/03

Go up one level in this thread


On October 20, 2003 at 15:00:23, Jeroen Noomen wrote:

>On October 20, 2003 at 14:52:09, Uri Blass wrote:
>
>Hi Uri,
>
>Sigh. You are really very stubborn. If even Alex and I together cannot convince
>you, then I simply quit. You would even question me if I would state that the
>earth is not flat :-).
>
>Simply letting a machine calculate for 13 plies on one position is not enough
>here. The loss is more than 30 plies in the future. Now be brave and take the
>Rebel or Fritz book and look it up. Or play some autoplay games after ... b4 and
>watch what happens.

Relex.

I do not claim that white is not winning and I even agree that white is probably
winning
I do not criticize your choice of the book but only the claim that black is
losing after b4 instead of claiming that black is probably losing.

The point is that it is not a position when I can be sure even in 99% that white
is winning without analyzing it for a very long time.

If the loss is more than 30 plies in the future and there is no singular line
then in order to prove the win you may need to analyze thousands of lines.

I could see in few lines that black is losing but it can convince me to give
probability of maybe 80% that black is already lost.

More lines can increase the probability but I guess that I will probably need
thousands of hours if I want to be sure that it is a win.

Of course it is better to have 100 lines that you are sure in 80% that they win
and not one line that you are sure in 99.9% that it wins so a decision not to
investigate if it wins after getting the opinion that it probably wins is
correct practical decision.

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.