Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Analysis, fortress draws, and hash tables

Author: Stuart Cracraft

Date: 21:53:06 08/07/04

Go up one level in this thread


On August 08, 2004 at 00:03:48, Robin Smith wrote:

>On August 07, 2004 at 23:49:26, Stuart Cracraft wrote:
>
>>I liked Zobrist and Kalme's research many years ago into this subject.
>>Kalme would provide a series of patterns to guide the program. Not specific
>>positions. Another example is David Wilkins program for tactical problems.
>>
>>Good to hear this kind of interest is still alive in this search-exhausted
>>brute-force selective-days. I am pro-patterns and pro-learning! But
>>the implementation is bedeviling.
>>
>>Stuart
>
>I can well imagine that "implementation is bedeviling" for "patterns to guide
>the program". That is why I am suggesting being able to enter "specific
>positions". Any while I have no doubt that for playing games this would be of no
>value what-so-ever, I believe that for anaysis it could be very powerful. Any
>time the analyst can confirm that a position the program says is +3.0, when it
>really should be 0.0, the analyst can plug the programs blind spot. In other
>cases, if the analyst wants to avoid some murky looking complications, he could
>enter a position evaluation of -10.0 for the position after entering the
>complications and see if there are any other clearer paths to an advantage. I
>can imagine many possible uses, but only for analyisis, not for improving engine
>play.

I see your point -- for guidance and learning on the human side rather
than for computer program improvement. Certainly an interesting idea.

>
>P.S. I noticed in you profile that you mention John Stanback. I met him at his
>home several years back, he lives in the same town as me (or at least he used
>to). A very pleasant fellow. Do you know if he still does any chess programming?
>
>-Robin

I wrote GNU Chess 1 and realized I wanted a better life so I searched
around and found John who was willing to donate a 7-year chess program
(as I recall -- this was 20 years ago.) This became version 2, 3, and
4. Later it seemed wise to go to bitboards and that's when version 5
came out.

John was a manager in an IC lab in HP in Ft. Collins, Colorado.
I don't know if he's retired or what nowadays. Always an interesting
commentator and programmer in the computer chess scene.

Stuart



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.