Author: martin fierz
Date: 02:19:40 02/14/03
Go up one level in this thread
On February 13, 2003 at 23:58:10, Robert Hyatt wrote: >On February 13, 2003 at 15:23:15, Peter McKenzie wrote: > >>Just wondering how many programs are using the Enhanced Tranposition Cutoffs >>algorithm, or if people have tried it and had trouble with it. I think I'll >>have a go at implementing it shortly. > >I tried it several years ago with no success. IE the cost was a wash with >the reduced tree size, but that was all. I had to write a special "Make()" >function that did nothing but update the hash signature so that I could >probe to see if the new position would produce a quick cutoff. Since there >was no gain (and no loss) I stopped using it. > >YMMV of course, and checker programs seem to report good results using this, >but their tree is a different shape. for my checkers program this is true - ETC costs me about 10% in nps, but typically reduces tree size around 20% on a think of a few seconds, so it's quite an improvement for me. i'm using ETC in nodes 3 or more ply away from the leaf nodes. checkers has a very low branching factor (something like 2), so as bob says, it's probably hard to compare to chess. cheers martin
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.