Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Position solved.

Author: Uri Blass

Date: 16:40:46 10/12/02

Go up one level in this thread


On October 12, 2002 at 18:38:58, Robin Smith wrote:

>On October 12, 2002 at 16:49:22, Hans van der Zijden wrote:
>
>>On October 11, 2002 at 20:54:54, Mike Byrne wrote:
>>Roland Pfisters program Patzer solved this position in 6'49'' on a Pentium 3 450
>>MHz almost a year ago. After 16 ply 1.Kc6 is found with an evaluation of -0.04.
>>
>>Hans.
>
>Hans,
>
>How does Patzer manage to solve this position?  The solution is an unusual
>fortress which involves putting black's king into a cage from which it cannot
>escape.  Does patzer have special knowlege for this type of position?  Do any
>other programs have this knowlege?  I would think no program would ever see a
>near draw score here.
>
>Robin

It is wrong to assume that programs cannot understand fortress positions.

Patzer understand some of them and David Omid's private program understand
more fortress positions.

Programs can understand everything that the programmers teach them.
Most of the programmers simply decided not to teach the program fortress
position because they did not consider them as important enough to waste
hundreds of hours in that problem.

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.