Author: Christophe Theron
Date: 12:32:31 01/04/98
Go up one level in this thread
On January 04, 1998 at 06:53:57, Thorsten Czub wrote:
>If M means MOVE this could explain why the main lines have a pattern on
>the base of 2 .
>This would correlate better !
>If you believe M means ply, than it is not obvious to me why the machine
>has pretty good main-lines, deep, after 1 second computing time and
>M=sero !
The deep lines have nothing to do with the full-width depth base, be it
M or M*2 plies.
The deep lines are a product of the selective search, which is conducted
beyond the full width search depth. The depth of this search seems to be
varying a lot, depending of what kind of position arise from the first
moves. If the position is not quiet at all (in check, threatened of
something...), the search extends, up to a quiet position, or up to 12
plies beyond nominal search (in Genius).
From what you and Don have found, the selective part is asymetrical
(pruning more of the computer's moves), the full width is not, unless
Thorsten finds a position denying this (using selective depth=0 to make
it very clear).
Also the selective part has a mechanism to decide if the position should
be extended or just evaluated without more search. There must be some
heuristic to prevent the selective search from growing into a
near-full-width search, which could happen if there are many threats for
both sides.
The selective part tends to end after a computer's move, which gives
line lengths that are a multiple or 2 PLUS 1(not "power of 2"). For
example you'll mostly see lines of length 3, 5, 7, 9, 11, ... Maybe
another asymetry here: threats against the program are extended, but the
program doesn't care very much about threats it can do on the opponent
(except checks, maybe).
The goal of the selective search seems to be: show that the position
reached by the full width search is safe for the program.
I'm not absolutely sure of all of this. This is data under discussion...
Christophe
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.