Author: Robert Hyatt
Date: 08:53:20 09/19/00
Go up one level in this thread
On September 19, 2000 at 11:41:33, Peter Fendrich wrote: >On September 19, 2000 at 10:22:03, Robert Hyatt wrote: > >>On September 18, 2000 at 17:42:06, odell hall wrote: >> >>>Hello CCC >>> >>> >>> How many think the new SSDF List is relatively Accurate? Personally I commend >>>SSDF for doing a outstanding Job, Based on my observations of the grandmaster >>>Challenge and other 40/2 events, I think the list is very reliable. >>>I believe it is safe to say that any top program running on a K62-450 is 2500 >>>elo, or very near. I think that now that the rating has been significantly >>>lowered, this list will be taken far more seriously in determining Fide rating >>>for Modern Programs. I am curious if some skeptics of the List in the Past, >>>consider the list still to high, or Just about right? Opinions Welcome >> >> >>For anyone that understands the statistics in the Elo system, the correction >>makes zero sense. There is no way to take two totally different rating pools, >>and adjust one set of ratings to make them comparable to the others. As has >>been pointed out so many times here, the _difference_ between two ratings is >>the _important_ piece of data. Not the raw ratings themselves. All this >>correction does is attempt to correct highly over-rated numbers on the SSDF >>list. Adjusting _all_ ratings on their list changes _nothing_ of course. And >>it still doesn't mean that their ratings have anything to do with FIDE ratings, >>other than the top programs are now a bit closer to "reality ratings"... > >Dr Elo couldn't even dream of how his rating system was going to be used. It is >designed for the old type of human tournaments and not for what is going on at >ICC or the thousands of games played by SSDF. There are no "reality ratings" in >this sense. I agree, although his formula will work for server chess. The only number that needs "tuning" is K. K=32 is fine for a few tournaments a year. It is _not_ so good for dozens of games a day. >Traditional human tournaments, matchplay, playing at ICC and comp-comp games by >SSDF means completely different conditions giving different results and ratings. >The elo system just don't overcome this. Completely correct, of course... Although the Elo system does work in any of the above. The "ratings" are not cross-compatible between events of different conditions. >The rating differences within the SSDF pool are comp-comp rating differences, >nothing more and nothing less. Everyone hopefylly agree with this... >I still think that there is a "cosmetic" point in adjusting the rating list so >that the top program ratings are close to what they probably would get in "old >style" tournaments. No one knows the right figures so why not a good guess? That is exactly what I said... They could also have added 2000 so that _nobody_ would confuse those with FIDE ratings, which would have been even better. Because now there are still overrated programs on the list, but also there are some that are likely well underrated now. When >adjusting the list, of cource all the ratings have to be adjusted by the same >amount - so nothing is really changed - no big deal. > >//Peter
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.