Author: Uri Blass
Date: 01:18:36 06/02/04
Go up one level in this thread
On June 01, 2004 at 12:55:37, Mike Byrne wrote: >On June 01, 2004 at 10:40:57, Uri Blass wrote: > >>On June 01, 2004 at 10:16:33, Drexel,Michael wrote: >> >>>On June 01, 2004 at 06:38:21, Uri Blass wrote: >>> >>>>On June 01, 2004 at 05:52:26, Drexel,Michael wrote: >>>> >>>>>On May 31, 2004 at 20:37:19, Mike Byrne wrote: >>>>> >>>>>>On May 31, 2004 at 19:27:20, Jonathan Lee wrote: >>>>>> >>>>>>>In the ICD chess software, the engines surpass 2750 and going over 2800 at 1.2 >>>>>>>GHZ? >>>>>>>How high can the Swedish rating system, SSDF, can it go? >>>>>>> >>>>>>>FIDE's and USCF's highest are 3000. >>>>>>>I learned later that ELO perhaps can go over 8000 (that would be somthing like >>>>>>>50+ ply for ELO). >>>>>>>Jonathan >>>>>> >>>>>>Without knowledge of the rating population, the absolute value of any ELO is >>>>>>valueless. The way SSDF is designed, a fixed pool of computer programs >>>>>>generally running and rated on older hardware with the new blood coming in on >>>>>>faster hardware with more modern programs - there is only one for the ratings to >>>>>>go and that is up. They have bad case of what I call the "Bloodgood" effect >>>>>>with the limited rating pool. >>>>> >>>>>Very true. SSDF and other computer rating lists are highly inflationary. >>>>>You can't compare it with FIDE rating system at all. >>>>> >>>>>There is not only a different pool of players. >>>>> >>>>>Chessbase of course is not unhappy with the present state. >>>>>After all the ratings are a good sales argument. >>>>> >>>>>Michael >>>> >>>>I do not think it is a good sales argument. >>>> >>>>Most players do not care much about the exact rating against humans when the >>>>program is better than them. >>> >>>That´s Nonsense. They _care_ about the strength compared to the top human >>>players. >>>The number of players who buy a new program in order to play against it at full >>>strength is negligible small anyway. >> >>I know that most people do not buy a new program in order to play against it but >>my point is that people buy a program because it is better than the previous >>version and the relative strength to the top humans is irrelevant. >> >>Uri > >Ratings , like it or not , are relevant to some consumers. They are used for >advertings purposes all the time - either SSDF or some pefromance ratings >against GM's etc. Ratings are to chess programs, is like Ghz is to PCs , top >cruising speed is to automobiles, etc. There are fair number of consumers who >believe they must always have the best, fastest, strongest, rarest etc. Just >ask Chessbase or just about any professional chess author - do they sell more >"Young Talents" or is their flagship products like Fritz, Shredder etc. > >What was the hottest chess program in the early 90 's - Chess Genius - why - >because it was considered the top back then. Which program do people want to >buy today - it is Shredder. Why because of knowledge people know that Shredder >is the strongest or they believe to be the strongest. Anybody who comes to CCC >is for the most part more knowledgeabe than the average consumer on the street >who knows nothing about Chess Program. If one of my chess friends tells me they >want to buy the strongest chess program, I would tell them right now I believe >it is Shredder 8. All of us here gave influence to a certain degree what chess >program our friends buy. Now as indidviduals , some of us may prefer one >program over another - which is fine - but collectively we cast our votes by >which programs we buy and which ones our friends buy. The people who sell Chess >programs know that "perceived" strength matters and ratings blend into that >concept of "perceived" strength. So to say ratings are irrelevant does not make >any sense to me at all. > >YMMV. I said that rating against humans are not relevant. I did not claim that rating are not relevant but they are relevant because people prefer a better program. The question what is the rating against humans is not relevant and if the ssdf list was 300 elo lower for all the programs the sales could probably be the same. 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.