Author: Bernhard Bauer
Date: 12:54:39 12/08/05
Go up one level in this thread
On December 08, 2005 at 14:39:24, George Tsavdaris wrote: >On December 08, 2005 at 09:05:23, Bernhard Bauer wrote: > >>On December 08, 2005 at 08:53:06, K. Burcham wrote: >> >>>On December 08, 2005 at 08:01:05, James T. Walker wrote: >>> >>>>Rybka still trails Fruitz 2.2.1 by 21 Elo even though it beat Fruit by 27 Elo in >>>>head to head match. Rybkas main defeat so far came at the hands of Fritz 8. >>>>All my games are run with each program haveing their own cpu/book (except Rybka >>>>plays without book) and ponder on.Rybka is a great program but should have it's >>>>own "official" book. >>>>Jim >>> >>>with no book, it has been stated many times here that one must take the square >>>root of pi, divided by x(a-c2)-3x(a2)over 3.1285. > >What the x and a and c2 and a2 represent? > >And how exactly is the formula? I mean what is the x(a-c2)-3x(a2) mean? > >Also what "over 3.1285" means? > > >>Once this value is applied to >>>your results you will see that Rybka has an Elo of 2875 and leads your Fritz 8 >>>by 22 games. > >How did you come up with this? > > >>Of course Uri will adjust this just a little when playing two >>>programs, one with book and one without. >>> >>>kburcham >> >>Actually your formula is only an approximation, though I have too agree, a good >>one. The true value is not 3.1285 but square_root( square_root(96) ). > >How did you come with 96? As you may know a chessboard has 64 places for 32 pieces. The sum is 96. Therefor. >Also what is the formula you are talking about and how is derived? You need a some mathematics. Actually some calculus mixed with statistics. Kolmogorov derived it for Botwinniks chess program in Sobolev spaces. > > >>However, that changes not too much. >>Kind regards >>Bernhard
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.