Author: Uri Blass
Date: 07:39:58 05/20/03
Go up one level in this thread
On May 20, 2003 at 10:25:22, Tim Foden wrote: >On May 20, 2003 at 09:39:42, Uri Blass wrote: > >>On May 20, 2003 at 02:45:47, Jorge Pichard wrote: >> >>> >>>Great progress by Movei, since I tested it last year. Congratulations Uri. >>>Please do NOT make an argument about being only ten games, this was only to get >>>an Idea of the progress of Uri program which I have not tested since last March >>>of 2002 against Chess Tiger 14.9 for Palm. >>> >>>PS: For those that constantly complaint about not testing enough, they should >>>concentrate their argument about those ICT3 short tournaments of only 9 games. >> >>I think that these games gave me some ideas to improve movei by adding rules >>when not to use my pruning rules but I will need to test it. >> >>Movei main problem seems to be in the endgame. > >I managed to watch some of the GLC-Movei games as they were played on FICS. One >of the things I noted was that in the early and middle game, movei was generally >searching a bit deeper than GLC. Towards the endgame GLC caught up, and as the >position got more and more simple, GLC searched relatively deeper and deeper, >ending about 4 to 6 ply ahead. I guess this is due to GLC using the hash table >information for cutoffs whereas Movei doesn't do this. > >Cheers, Tim. Yes I also suspect that the same productive pruning rules that I have in the middle game are not productive in the endgame so my first try may be simply to tell movei not to use one of it's pruning rule when the total value of pieces is not bigger than 18 so the result may be that movei searches even less deep in the endgame but I prefer that it play better and not that it earches deeper. Uri
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.