RE: [Bug-gnubg] [Fwd: gnubg search function]

[Albert Silver]

I have the same feeling, plus many of the techniques are clearly only
usable by chess such as mate extensions, etc. But what about others?
Mini-max is one of the oldest and most basic in chess, so Joern's
mention of its use in GNU suggests there might be others. I can’t say
how they would be used in a backgammon search routine but perhaps
something like null-move could. Null-move was a revolution in chess
engines, and there are others. Here are a few links describing the
techniques for any who are curious.

Bruce Moreland's site
http://www.seanet.com/~brucemo/topics/topics.htm

Ernst Heinz's site
http://supertech.lcs.mit.edu/~heinz/dt/

Paul Verhelst's site
http://www.xs4all.nl/~verhelst/chess/search.html

and even a description of MTD(f) an improved mini-max algorithm
http://theory.lcs.mit.edu/~plaat/mtdf.html

There are many others of course.

                                        Albert

> -----Original Message-----
> From: Jim Segrave [mailto:address@hidden On Behalf Of
Jim
> Segrave
> Sent: Thursday, March 27, 2003 10:16 AM
> To: Albert Silver
> Cc: 'Øystein Johansen'; address@hidden; 'bug-gnubg'
> Subject: Re: [Bug-gnubg] [Fwd: gnubg search function]
> 
> On Thu 27 Mar 2003 (10:03 -0300), Albert Silver wrote:
> > Is this something of your invention or are did you find a way to
adapt
> > one of the many documented search techniques used in chess engines?
Just
> > curious. In any case if it really works we may actually see GNU
> > *playing* 4-ply. Now that would be amazing.
> >
> >                                             Albert
> 
> I have a suspicion, but could well be wrong that most of the pruning
> techniques used in chess programs aren't going to work so well with
> backgammon, because the choice of replies to a move is so conditioned
> by the possible rolls. But I'd love to be proved wrong.
> 
> --
> Jim Segrave           address@hidden