Author: Uri Blass
Date: 21:48:14 01/29/06
Go up one level in this thread
On January 29, 2006 at 21:42:51, Djordje Vidanovic wrote: >On January 29, 2006 at 17:51:29, George Tsavdaris wrote: > >>On January 29, 2006 at 17:08:49, Djordje Vidanovic wrote: >> >>>On January 29, 2006 at 11:46:09, George Tsavdaris wrote: >>> >>>>On January 28, 2006 at 21:19:05, Djordje Vidanovic wrote: >>>> >>>>>A couple of days ago I promised to report on the long match I started between >>>>>Rybka 1.0 Beta 13 and Fritz 9. Conditions: game in 10 min plus 2 sec increment, >>>>>same home-cooked book with drawish lines, hash both 128 meg, ponder on, dual >>>>>Xeon 2.66, 3-4-5 men egtb's. The current (and final) score after 214 games is: >>>>> >>>>> >>>>>Home_Machine, Blitz:10'+ 2" >>>>> >>>>> >>>>>1 Rybka 1.01 Beta 13 +112/-32/=70 68.69 147.0/214 >>>>>2 Fritz 9 +32/-112/=70 31.31 67.0/214 >>>>> >>>>> >>>>>this points at about +135 Elo in favour of Rybka. An amazing result. >>>> >>>> Not amazing, but super amazing! If this success continues at longer time >>>>controls then what can i say?!? Do you think that it will or it will normally >>>>decrease a bit.....? Also was this beta 13b or 13.....? >>>>Can you send me the games to: gtsavdar@auth.gr ? >>>> >>>> >>>>> >>>>> >>>>>Djordje >>> >>> >>>This was the 32 bit Rybka, even though the first, shorter, part of the match >>>(about 60 games) was played with the 64 bit Rybka and ponder off on an AMD64 @ >>>2.4 Ghz. >>> >>>Jorgos, I am afraid that I wouldn't feel too happy about sending you the games >>>as many of the lines in the book will eventually be part of a competitive >>>opening book :-( >>> >>>I hope you will understand my reluctance to share the games. >>> >> >>OK no problem although i wanted to see the games....:-( >>Can i ask for which program do you create the book....? Also it seems a bit >>strange that you are making an opening book for an engine, a book that is based >>on drawish lines?!?! I don't understand.... > > >Drawish lines go to a certain depth of course, securing a solid position. Then >they branch out and I try to pick the branch that might lead to a better >position according to the assessment of the program I am trying to make the >openng book for. I do not understand why drawish lines. If you want to win games it is more logical to have winning correct lines when the winner did no mistakes. If the opponent blunder into the trap you win. If the opponent does not blunder into the trap and get you out of book you still have a game that you have chances to win. Note that of course book is not symmetric and if you have winning line in book it does not mean that you have losing line if you are in the second colour inside the book. 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.