Computer Chess Club Archives


Search

Terms

Messages

Subject: Limiting the QSearch

Author: John Scalo

Date: 10:15:25 12/30/97


Maybe it's because I don't have any sort of SEE, only a



<small takes big>/<equal exchange>/<big takes small>



ordering, but I see a *huge* explosion in my qsearch which is captures
only (no checks). I regularly hit depths of 20+ not including plies in
the regular search. First off, I'm wondering if the pros around here hit
those depths in their qsearch?



I'm also wondering if anyone has experimented with limiting the qsearch.
After examining the qsearch dives to ply 10 and beyond, I found that the
exchanges are just garbage combinations that would never get played. So
I limited my qsearch to a depth of five ply and don't consider "big
takes small" captures at the final ply, since those will most likely be
bad moves. In this scheme, the program is three or four times faster and
in initial testing doesn't appear to be any worse in strength!



Has anyone else tried this? What are the possible negative effects
besides the occasional tactical miss?



Thanks!

-j



This page took 0.01 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.