Author: Charles Roberson
Date: 20:25:14 09/06/04
Go up one level in this thread
It doesn't seem to slow the overall program because you see the mate
one ply earlier. This can give more prunning and more stability from
ply to ply. After implementing it (about a year ago), my program produced
better numbers (time and nodes) for all my benchmark suites.
My assumptions in my previous post was that a full ply extend is made
for checking moves.
BTW: Based on my tests, a major missing feature of my engine is that I
have not implemented single reply extensions. And my definition is that
single reply does not have to be preceeded by a check.
The other major problem for me is a bug in 3-fold rep of position.
I thank Leo D and Gunther S for pointing it out to me.
Charles
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.