Author: Uri Blass
Date: 10:56:01 12/30/03
Go up one level in this thread
On December 30, 2003 at 11:49:11, Christophe Theron wrote: >On December 29, 2003 at 12:36:01, Daniel Clausen wrote: > >>On December 29, 2003 at 11:27:40, Christophe Theron wrote: >> >>[snip] >> >>>I am less and less interested in participating in tournaments. They are just >>>lotteries. >>> >>>I think that playing more games (and using faster time controls so playing a lot >>>of games becomes possible) would be good, and I could then reconsider. >> >>I think the problem is not really that tournaments are lotteries. Afterall, they >>always were, to a certain extend. And not just in sports, everywhere. > > > >Which does not add anything to their scientific value. Only to their >psychological importance maybe. > > > > >> I think >>the problem is that the results of these tournaments are so important. > > > >The fact that people give them so much importance shows how misinformed the >general audience is. Even a more specialized audience like the CCC readers are >still badly misinformed, even after years of education on the subject. > > > > >> If people >>were less subjective but more objective, they would see these tournaments as >>fun-events and when it comes to real strength they would consider a rating list >>which includes _a lot_ of games. (although, as Vincent posted recently, the SSDF >>probably also has their own problems (I don't know whether it's true or not >>since I'm not involved at all in this, but it sounds likely to me that some of >>his claims are valid there)) > > > >There is no comparison between the result of any tournament and the SSDF list. > >The SSDF list is light years ahead. It is highly valuable. I think that other tournaments give information that the ssdf does not give. in comparison between WBEC and SSDF I find some advantage for WBEC. 1)All games are available. 2)No autoplayer that is a possible source for errors. 3)only 4 games between every 2 opponents so learning is relatively less important. Today there are 18 programs in the highest divisions that means 68 games for a program. I know that it means that the statistical error is clearly bigger than the ssdf but it is still smaller than world championship and you can expect 2-3 tournament in a year so you are going to get some hundreds of games in a few years and it is possible to do comparison between commercial programs. You can send new tiger to every tournament because updates for the next tournament are allowed so you do not get hundreds of games of the same tiger but the same is for other programs and I think that comparison between results of all Tigers and result of all Sjengs or all Ruffian in 2 years may be interesting. number of games is enough for you to be practically sure that the top commercial programs are not going to go down from the premier division to a lower division in the next 2 years unless I underestimate the progress that is going to be done in the next 2 years by a lot of amateurs. 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.