Author: Jeremiah Penery
Date: 02:44:38 11/29/00
Go up one level in this thread
On November 29, 2000 at 04:20:04, Uri Blass wrote: >On November 29, 2000 at 03:05:50, Jeremiah Penery wrote: ><snipped> >>I have similar results with my modified Crafty. I'm still working on teaching >>it to know how and when to attack, and how much the attack is worth...I think >>the default Crafty may take ages to find this, because it doesn't value white's >>attack until he begins to regain the lost material. > >The question is if your modified crafty is better in games than the original >Crafty. >A possible problem that I can imagine is that it may go for wrong sacrifices. > >I guess that if you prove that it is better then Bob will be happy to use it. I'm trying. :) Part of the problem is that it's probably being tuned for my slow PII-300, so I have no idea how it would react on a faster machine. But I have been working on this 'version' for quite some time. Playing previous incarnations of it (less-tuned ones and such) on FICS/ICC, they did value the attack far too much. Hopefully I've solved at least that problem, and now maybe it values the attacks too low sometimes (in cases where it clearly has a winning attack, but evaluates it low). To solve that I need to teach it more about how to attack (number and type of attacking pieces, king shelter, and number and type of defending pieces).
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.