Author: Dann Corbit
Date: 13:21:41 09/28/99
Go up one level in this thread
On September 28, 1999 at 16:09:54, walter irvin wrote: >i think ssdf should only use ratings on a single mhz speed . if they did that it >to me would be a lot better . In what way? The slower machines are what the tests used to be run on. They have a limited number of faster machines. Are you planning to donate some faster hardware? I am quite sure they will accept, unless they come with some strings attached. >if all they got is 200 mhz then so be it .i think >it is stupid not to rate cm 6000 at 450 mhz when it still had the highest rating >at 200 mhz .why rate any of them at 450 .for all anyone knows rebel might be the >strongest at 600 mhz ,or mchess or crafty .i think all the programs should be >rated on ave hardware .all of them ,even the ones that you see on the list at 90 >mhz .pull them all into line then you can say this one is this strong that one >is that strong .maybe it would require a lot of work , but it would leave no >doubt as to where the programs stand and just exactly how much progress there >has actually been .its stupid when a new program comes out so you also give it >the latest hardware advantage .put them all on equal hardware , if a program >realy is better it will climb its way to the top , with out all the extra >advantages!!!!! Go ahead and create your own testing group. Make sure that the programs are set up according to carefully controlled and known positions. Also, make sure that your mathematics are correct for all of your calculations. Then, run a few thousand games through this setup and publish your results. I suggest you get a few dozen 500 MHz PIII's since those will be pretty much the minimum acceptable machine by the time your tests are through, if you can find enough volunteers. Of course, in a few months those 500 MHz machines will definitely be passe and you will be roundly criticized for running those crappy tests on those crappy machines. But _then_ if you get some faster machines -- will you wait 6 years until you have had an opportunity to rerun each and every test or will you report the results as they come along. [FCOL] People who do not understand seem to have no difficulty in being critical. I suppose that is the Usenet culture in general for you.
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.