Author: Uri Blass
Date: 06:07:28 05/14/05
Go up one level in this thread
On May 14, 2005 at 06:17:46, pavel wrote: >On May 14, 2005 at 02:33:26, Dann Corbit wrote: > >>On May 13, 2005 at 20:57:53, pavel wrote: >> >>>5/13/2005 8:49:06 PM : >>> >>> Program Elo + - Games Score Av.Op. Draws >>> >>> 1 Aristarch 4.50 : 2508 6 6 11358 52.2 % 2492 25.2 % >>> 2 Yace Paderborn : 2492 6 6 11358 47.8 % 2508 25.2 % >> >>I would like to get the games, if possible. >> >>Amazing that the chosen opponents are so closely matched. Even now, after 11K+ >>games, it is only just now barely proven that Aristarch is stronger with the >>configuration and time control that you chose. >> >>This is definitely the most precisely that ability has ever been scaled. >>Congratulations on a fascinating experiment. >> >>Is it at the final now, or are you going to run even more games? > >Thanks Dann. > >No, it is not final. My primary goal was to reach 10k games, now my goal is to >play as long as I don't get bored or I reach 0 error margin. Whichever comes >first. > >The size of the PGN file (stripped off CB game scores and rarred) is about 3mb. I can only say that it is a bad compression job and it is probably possible to compress it to 1 mb. My calculation is as follow: 1)It is no problem to have average of 6 bits per move(you only need legal move generator and remember the number of the move and usually the number of the legal moves is less than 64 and to compensate the few cases when the number of moves is more than 64 you have a lot of cases when the number of moves is less than 32 and you need less than 5 bits for a move. 2)a game has average of less than 133 plies 3)133*6M<800 bits per game=100 bytes per game 4)100 bytes per game*11358 games is only slighly more than 1 mbyte. Uri
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.