Author: Dann Corbit
Date: 18:39:42 05/05/05
Go up one level in this thread
On May 05, 2005 at 21:16:11, Dann Corbit wrote: >On May 05, 2005 at 21:08:25, Marc D wrote: > >>Hi >> >>Would it be of interest doing Shuffle Chess Tournaments with selected >>Engines and post the results here? > >It would be interesting. > >>It would show how good these engines are without opening book and >>just plain chess knowledge. > >I doubt if that is what it would show. > >>So let me see your thoughts about it > >I think it would be better to play computers against stron human shuffle chess >players for a while in order to create an opening book. Before that happens, >you will see idiotic repetitions where a computer program loses the same opening >over and over and over. > >Why is it (for instance) that humans study opening theory? Because many opening >sequences are decidedly non-obvious and have had the wrinkles ironed out by >thousands upon thousands of attempts. So, eventually, we discover what is >playable and what is not playable. > >Here are the very humble beginnings of an opening book for Chess960: >ftp://cap.connx.com/pub/chess-engines/new-approach/fc960.epd > >You will notice that it does not even have one full ply yet. Eventually, it >will be useful. For now, it is just a curiosity. Stats: min(seconds) max(seconds) ave_seconds min(ply) max(ply) ave_ply min(eval) max(eval) ave_ce min(nps) max(nps) ave_nps min(nodes) max(nodes) ave_nodes 504 2313 878.86 19 23 21.36 2 43 17 411199 490393 446975 219867942 1053609055 392733936
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.