Author: Uri Blass
Date: 00:47:40 09/12/01
Go up one level in this thread
On September 12, 2001 at 01:54:40, Sandro Necchi wrote: >On September 09, 2001 at 17:01:37, Uri Blass wrote: > >>On September 09, 2001 at 15:48:43, Gian-Carlo Pascutto wrote: >> >>>On September 09, 2001 at 15:39:35, Uri Blass wrote: >>> >>>>Programmers consider the ssdf as more important than championship including >>>>christhophe(the programmer of tiger) >>>> >>>>The most important competition is the competition that is considered by most of >>>>the programmers as the most important competition. >>> >>>We need a programmers-only poll. > >Well, how do you know they will tell you what they really think rather than what >is more convenient to? >>> >>>The reason why some prefer SSDF is probably because they are doing well >>>there. Also, SSDF is available the whole year. The WMCC results are >>>usually soon forgotten. I guess the best is what has the best effect >>>on your sales though. >>> >>>>All the top programs participate including Shredder and Shredder lost matches >>>>against Deep Fritz and Tiger. > >You are comparing 2 different things. In the SSDF the programs have a wider >opening book and they are left with their learning system. >In the Championship the openings are restricted and the selection is made/can be >made on the opponent. I know and I guess that you simply did better opening preperation in the championships. I do not use the restricted opening of you in my correspondence games so they are not relevant for me. >It is like to put Ferrari and Fiat cars on the same level. > >>>I can name at least one program that has won several titles that >>>has never been in SSDF. >>> >>>-- >>>GCP >> >>Shredder was in the ssdf(Shredder2). >>Shredder is also going to be rated by the ssdf(shredder5.32) >>Shredder4 was hiding under the name 204. >> >>I remember that shredder4 could get >>only 4th place when Fritz6 Junior6 and tiger12 were number 1,2 and 3 >>respectively. > >The learning feature in the SSDF can be as good as 50 points. So a better >learning can make the difference. I believe that all the chessbase engines use the same learning algorithm. >Fritz and Deep Fritz have a opening book that has been refined for some years, >so this helps the learning even indirectly. >Shredder may loose game to find the best opening or to avoid the weak ones. at >the end this cost points! I do not understand the reason that it did not happen with shredder2 (I count only the first games of shredder2 in matches because I know that shredder2 had no learning). I know that shredder2 even could lead some matches against top programs with better hardware(probably thanks to a better book that was optimized for shredder because I do not believe that Shredder2 was really better than the new programs). If shredder5.32 is a clearly better engine than shredder2 than you could use for shredder5.32 the same opening book of shredder2 and start the matches with even better results than Shredder2 so the excuse of a bad opening book does not seem to be convincing. Uri
This page took 0.01 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.