Author: Jeremiah Penery
Date: 19:15:53 10/02/03
Go up one level in this thread
On October 02, 2003 at 20:49:53, Dann Corbit wrote: >On October 02, 2003 at 20:15:25, Jeremiah Penery wrote: > >>On October 02, 2003 at 16:00:23, Dann Corbit wrote: >> >>>Interesting is the large number of nodes the new 19.4 version examines. About 3 >>>times as many as the MPC version. >> >>About that MPC stuff in Crafty - I tested it some, but the program almost always >>generated nonsensical lines tending toward a drawscore, even when there was an >>easy tactical win. When I removed the code, these lines were seen as normal. >> >>Have you done any testing to verify the correctness of the MPC code in those >>versions of Crafty that you have? > >I ran a short contest and the MPC version performed better. >Maybe it does the pm correctly but clobbers the pv. It had pm and score completely wrong as well in the tests I did with it. I'm interested to see you test it on some more tactical positions. There was one analysis in particular that I was having trouble on, I'll see if I can find it in the Archives or something. It's possible that the MPC code was interacting badly with some other code I have inserted into Crafty. I didn't really check it out too much.
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.