Computer Chess Club Archives




Subject: Re: Static Exchange Evaluation (SEE) for pruning in quiescence (?)

Author: Robert Hyatt

Date: 09:39:22 08/19/03

Go up one level in this thread

On August 19, 2003 at 03:03:08, Peter Kappler wrote:

>On August 19, 2003 at 02:48:36, Omid David Tabibi wrote:
>>It seems that some programs use Static Exchange Evaluation in order to prune
>>losing captures in quiescence search. In the following position, an SEE will
>>deem the move 1.Rxd7 a losing capture, and it might get pruned in quiescence.
>>However, in fact this is a winning capture since Qf6 is attacked after 1.Rxd7.
>>How do such programs solve these kind of problems?
>>[D]3r2k1/pppb2pp/5q2/5p2/3R1P2/2B5/PPP3PP/5RK1 w - - 0 1
>Rxd7 should just be a draw after Qb6+ Bd4 Qxd4 Rxd4 Rxd4.  Or maybe Black can
>play for a win with Qb6+ Bd4 c5 Rxd8 Qxd8 Bxc5 Qc7.
>Sorry, I realize this is all beside the point.  Couldn't resist.  :)

But it makes an important point.  Anybody that thinks the q-search is "very
accurate" is living in never-never-land.  When you limit a search to just
captures (or just captures and checks) that is a _severe_ limit.  Try to play
a game where for each of your possible moves, you _only_ consider checks or
captures by your opponent.  You get killed.  And easily.

Worrying about the q-search is hopeless.  The main point is to make the rest
of the search as accurate as possible, ditto for the evaulation, and then hope
those two things cover up the inaccuracies in the q-search itself.

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.