Author: Harald Faber
Date: 01:45:59 08/20/02
Go up one level in this thread
On August 20, 2002 at 02:14:24, Jouni Uski wrote: >Very interesting! BUT I fear we need as usually many 100s of games to get any >reliable data :-( > >Jouni To get staistical relevance, yes, there is no way than play at least 500 or better 1000 games each program. *I* won't do that, takes FAR too long. BUT I will get an <i>idea</i> in which direction to think. And I will repeat the openings after each run so that direct comparison is possible. Notice that this is the first time this kind of test is done. Unbelievable. Seems as if "Best-for-Fritz" and this kind of "tests" (I won't write it down what my word for these tests is...) finally have won... maybe time to leave computerchess soon. The translation into English is complete now and will be online late this evening or tomorrow morning. Have fun.
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.