Author: Uri Blass
Date: 13:12:43 10/11/00
Go up one level in this thread
On October 11, 2000 at 13:35:56, Timothy J. Frohlick wrote: >Jim, > >Gambit Tiger 1.0 on a PII 333 with 64M RAM does not get the solution either >after 8 minutes. I do believe that in order to get this solution the programs >would have to see the consequences of the "fifty move rule" and search fifty >moves ahead or more. Deeper Blue would have seen it but none of todays PCs can >do this. How do you know that Deeper blue could see it I believe that it cannot see it. Programs do not need to search fifty moves ahead or more in order to avoid Nxb6+ because of 2 reasons: 1)It may be enough to see that another move give better evaluation without seeing that Nxb6 is a draw 2)It may be enough to see by evaluation and not by search that Nxb6 is a draw. The programs can't conceptualize yet and the knowledge for the >millions of specific exceptional positions like this would slow the programs to >a crawl. Knowledge will slow programs only if they use the knowledge. Programs may use the knowledge for the first 0.1% of the time and stop using it later if they find that the knowledge is not relevant. I believe that the real reason that programs do not know about this position is the fact that programmers have more important things to do in order to do their programs stronger. 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.