Author: Günther Simon
Date: 02:11:28 09/03/05
Go up one level in this thread
On September 02, 2005 at 22:15:11, Jorge Pichard wrote: >On September 02, 2005 at 14:20:13, José Carlos wrote: > >>On September 02, 2005 at 12:34:12, Jorge Pichard wrote: >> >>>On September 02, 2005 at 09:21:02, Kurt Utzinger wrote: >>> >>>>On September 02, 2005 at 09:17:54, Harry Schnapp wrote: >>>> >>>>> Hi Kurt >>>>> Most games were practically not played with my book !!!!!!!!!!!! >>>>> Only the first 4-5-6 moves!!!!!!!. >>>>> Harry >>>> >>>> >>>>Hi Harry >>>>In that case we indeed have the proof that the games >>>>are worthless. Thanks for your information and please >>>>give Jorge Pichard an advice how to set up Arena correctly >>>>for the use of specific books -:) >>>>Kurt [http://www.utzingerk.com] >>> >>> >>>As long as both programs are using the same setup the the strength of the >>>program either with books or without books is shown with these 40 games. >> >> This statement is clearly false in the general case. Take a starting position >>with black queen taken out, all of the other pieces in the starting position. >>White will always win no matter what programs you match, so both program will >>seem to be equally strong. Same setup does not imply fair conditions to measure >>strenght. Own book for each program do. >> >> José C. >> > >Jose the same will happen if you take the White Queen out, of course Black will >win all the time. When I said the same setup, I meant to say that Black and >White play with the exact setup, using the same book, as in this case the >Mainbook of Harry. No Harry said you don't use it ;) He offered you his help to set it up. You seem still not to understand, what was prayed for years here. If the book or whatever you use is bad, the tendency of the result always goes towards 50% (if both change colours or the book contains _always_ very bad lines) and towards sheer randomness in the case of random book selection and _not always_ having bad lines, thus you measure nothing...and by no means it has something to do with 'fairness', the result simply is _not_ correlated with the programs. You just measure the worseness of the book, (= bad lines/number of games), if you play enough games. It can't be that hard to understand this, think about. >Also if both programs play without Opening Books, no program >has any advantage as long as they alternate white and black after each game. That is true, but has nothing to do with above scenario. Guenther > >Jorge >> >>>PS: Take it however you want to take it, you can test Spikes 1.0a vs Toga II >>>without Books if you want. Your test was done with version 1.0 Mainz NOT with >>>version 1.0a.
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.