Author: Robert Hyatt
Date: 06:01:38 01/08/99
Go up one level in this thread
On January 08, 1999 at 00:54:58, Don Dailey wrote: >On January 08, 1999 at 00:12:22, Robert Hyatt wrote: > >>On January 08, 1999 at 00:10:33, Robert Hyatt wrote: >> >>>On January 07, 1999 at 19:50:47, Marc wrote: >>> >>>>OOps, I don't think they're called ECT, I forgot the acronym. >>>> >>>>Enhanced Transposition Table Lookups, I think. >>> >>> >>>I'm not sure what this is, can you elaborate? >> >>I think I remember this now. The idea is that whenever you get to a ply and >>you don't have a good hash move to look at first, try each of the moves, compute >>the new hash key, and see if any of those lead to a hash hit. If so, try that >>move first... >> >>I never had a lot of good luck with this, but I didn't try real hard >>either. you need a quick piece of code to compute the new hash keys without >>doing a lot of unnecessary work as well (ie don't use makemove() to do this) > > >Same story with me. I did try it, but I didn't try too hard and didn't >find it particularly good or bad. > >- Don There may be "something" there. Jonathan has reported very good results in his checkers program Chinook. My problem was that I used MakeMove() to compute the hash keys for the probes. I tried to "estimate" what that cost was, because ETC did reduce the size of the tree some. But when I factored everything in, it didn't help or hurt. And since I try to adhere to the KISS principle, I took it out. It is worth another check sometime however...
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.