Author: Joshua Lee
Date: 09:58:34 05/13/01
Go up one level in this thread
I would agree to the ratings in the SSDF list to being too High but by over 200 elo points i can't agree with if you want proof go and look at the ratings against humans for the few programs ofcourse before they lowered the list by 100points which was good of them. Take the ratings and compare, some programs lost points and some gained so you could say add some to the list but you can also lower the list according to the results which if i remember correctly it was +- anywhere from less than 10 points all the way to 200 or more so common sense would tell you that there Are Not Enough Games Against Humans!!! Sorry but every time i try explaining this to the SSDF people you deny or you try and convince us that your statistics are foolproof when in reality they are so bad you might as well not publish a list unless you have 500 or more games against humans. Look every program that draws up a rating list will tell you that , for example go create a elo list from your ssdf games and include the games against humans it should tell you that there are NOT enough games what the games you do have only show how the programs that played against the humans will do against humans. Think about this if you want to better the list take your list of computers not on the SSDF list lower or raise what performance they would have if on the same hardware and you might just see a better list .... C'mon there is No way Junior 6 on Hardware 9.17 times slower Can POSSIBLY be only 109 points away from Deep Junior on a 8way 700Mhz system and even furthur so considering you have no idea how much memory they used for hash tables etc. I read somewhere that performace ratings are 180pts higher but i am not sure about that. in that case the list would be right
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.