Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Thanks for the info (Was: More information + a couple of diagrams)

Author: Dave Gomboc

Date: 09:52:56 10/08/00

Go up one level in this thread


On October 08, 2000 at 11:06:24, Heiner Marxen wrote:

>On October 08, 2000 at 10:45:45, Dave Gomboc wrote:
>
>>On October 08, 2000 at 10:36:14, Heiner Marxen wrote:
>>
>>>On October 08, 2000 at 02:00:25, Dave Gomboc wrote:
>>>
>>>>In fact, there's an entire Ph.D. thesis on it.
>>>
>>>Could you please give a more detailed hint / reference ?
>>>I have thought about this method myself, but expect a Ph.D. to
>>>contain even more interesting results.  :-)
>>>
>>>Heiner
>>
>>Campbell, Murray S.  Chunking for Abstraction.  Ph.D. Thesis, Carnegie-Mellon
>>University, 1989.
>>
>>I don't know how easy it is to get hold of it -- it should be (relatively)
>>easier to get a copy of Computer Chess Compendium and check out the chapter on
>>coordinate squares there.
>>
>>Dave
>
>Thanks very much for the quick and accurate answer!
>
>I will try to get hold of both.
>I expect the Compendium to contain a sort of overview of the thesis,
>not a copy of the thesis, right?
>Since I am interested in all details, I will probably need the thesis.
>
>I will try to use this corresponding squares technique in Chest to
>prove (sometimes) that a position is a draw, and therefore cannot be
>won in any depth.  Up to now, the usual search techniques cannot do this.
>
>For a stipulation "white plays and holds draw" I do not yet have any
>method to construct a positive answer.
>
>Of course, I do not yet know whether there will be a "practical"
>implementation (practical == fast enough), but I would like to try.
>
>Heiner

I misremembered the thesis title.  It is (if my memory serves me correctly this
time! :-) Chunking as an Abstraction Mechanism.  The article in Computer Chess
Compendium is not the same!  It is a different article on co-ordinate squares --
I believe it was Kenneth Church's B.Sc. thesis.

You should be able to prove with look-ahead that "white to play and draw" is
feasible, if you look ahead far enough. <grin>

Dave



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.