Re: [Bug-gnubg] [Q] Full-width search

[Thomas Hauk]

On Fri, 12 Dec 2003, Joern Thyssen wrote:
> The move filter looks fine. Can you send some examples where
> FindBestMove disagree with your code?

I have an uncanny ability to answer my own question after I've asked it, 
but not a second before.

Upon further reflection, I believe the difference between the two searches 
comes from the fact that all the moves at depth=1 are being scored -2.000, 
a gammon loss, and since I generate moves in a different manner, and sort 
them in a different manner (actually Gnubg doesn't sort moves at all), 
it's possible that my searches get a different move chosen.

> In fact, you can check this yourself if you are able to setup the
> position in the GUI. Press "hint", mark all moves, and press "1" (for
> depth 3).

I don't use the GUI... I write my own drivers and call functions directly. 
Close to the metal. 8)

BTW... while I am here... I have the same problem compiling gnubg on 
another machine as a previous poster has, involving having versions of 
automake, aclocal, etc, which are "too new". I followed Achim's advice and 
it didn't help. If anyone would like to take the compilation discussion 
"offline" from the mailing list, I'm interested in getting the problem 
solved.

--Tom

-- 
Some people, when confronted with a problem, think "I know, I'll use 
regular expressions."  Now they have two problems.
  --Jamie Zawinski, in comp.lang.emacs