Computer Chess Club Archives


Search

Terms

Messages

Subject: practical and theoretical opinions sought about multicut ala Schaeffer

Author: scott farrell

Date: 17:52:19 09/06/03


I am implementing multi as per schaeffer, see :
http://www.google.com.au/search?hl=en&lr=lang_en&ie=UTF-8&oe=UTF-8&q=+site:www.cs.ualberta.ca+brute+force+selective+extensions+search
http://www.cs.ualberta.ca/~yngvi/Papers/tcs2001.pdf

The idea is, when you expect a cut-node (to fail-high?) (ie. in PVS
alpha!=beta-1), search the first 12 nodes at r-2, if you get 3 fail-high, then
fail high straight away, as probably one of them would have failed high at the
real depth.

Has anyone tried this? use it currently?

In my implementation, I try multicut before doMove when alpha!=beta-1 and not
root etc, if it suceeds I fail-high without doing doMove and searching at the
current depth.

Scott



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.