Author: William Bryant
Date: 13:38:35 04/04/99
Go up one level in this thread
On April 04, 1999 at 11:59:40, Christophe Theron wrote: >On April 03, 1999 at 18:30:23, William Bryant wrote: > >>Are the results obtained on the LCTII depth dependent? > >Is the strength of a program depth dependant? > >Yes, of course. > >The combinations in the LCT-2 are quite easy, and depth 6-9 is generally enough. >Many programs of 300MHz processors will solve them in less than 30 seconds >(except CMB11 and CMB12, which are almost unsolvable). > >I have found that many of the positional problems need more depth (often 9-11). > >In the endgame positions you'll need both depth and a lot of knowledge >(especially for FIN08 and FIN09). > > >>I was rather >>disappointed in the initial score I obtained on the LCTII (score was 2120). I >>don't have Null Moves and was searching 8 to 10 ply on the positional tests, 7 >>to 10 ply on the tactical tests, and 11 to 17 ply on the endgame tests. >> >>Is this an indication that the eval needs more work (which of course it does), >>or that Null moves, which should increase the search depth, are more likely to >>increase the strength of the program. > >Both. > >2120 without null move is already nice (on which processor?). Do you have some >kind of selectivity? > > > Christophe This is on a G3 - 266mhz (an 'old' G3). No selectivity yet. Just Alpha-Beta, Aspiration windows, PVS, Move Ordering. Null Moves are next, then egtb's, then opening book with an improvement in the eval along the way. William wbryant@ix.netcom.com
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.