Author: Derek Paquette
Date: 15:29:25 09/03/04
Go up one level in this thread
On September 03, 2004 at 17:48:41, Peter Skinner wrote: >On September 03, 2004 at 17:17:04, Robert Hyatt wrote: > >>On September 03, 2004 at 16:56:53, Peter Skinner wrote: >> >>>On September 03, 2004 at 16:53:43, Robert Hyatt wrote: >>> >>>>On September 03, 2004 at 16:12:57, Eran Karu wrote: >>>> >>>>>. >>>> >>>> >>>>2 >>>> >>>>look at main.c comments. >>> >>>I find the pawn changes to be much better than v19.15. This is from about 60 >>>games thus far ar 30/0 vs Ruffian 2.1.0. >>> >>>Peter >> >> >>There was definitely a long-term problem. The blocked pawn code was doubling up >>behind the isolated pawn code in ways I didn't intend, making the penalty much >>too high in some cases. Was watching a game the other night where the score was >>about +1.0 yet it looked only marginally better for white to me. Investigation >>found the problem and some immediately better pawn structure play in the games >>after the fix.. > >I have noticed the same as well. The pawn structures that it now creates seem to >be much more solid, and long term based. Before it was almost like Crafty wanted >to swap out the pawns quickly to get to the "meat" of the game. > >Peter It is possible crafty might fall into the same trap as shredder when it comes to playing humans? it just doesn't open the board up now? CT15 (anti human) would open the board up, even if it lost 0.40 of a pawn, rediculous vs a program but it worked well vs humans as we saw in Argentina might crafty be taking the same ill-fated trip when it comes to humans with this 'change' ?
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.