Author: Ulrich Tuerke
Date: 04:48:10 05/10/01
Go up one level in this thread
>... > >I don't think it is a very strong correlation though :-) . > >Tim asked a question about the test specification - the test offers three >different rating schemes to scale the results : USCF , Swedish and British - I >was simply stating fact . I didn't post the reference to the USCF rating in the >result list - there are simply numbers posted : the more the better ( that's >perfectly valid ) . > >>I don't buy that. >> >>Let's call it "Kaufman rating" or what ever. >> >>Uli > >I think this test is interesting - the solutions given seem to stand the test of >time. I agree. The test certainly measures some kind of ability of the programs and the numbers are not quite useless. However I'm afraid, the tactics part of it is a bit outdated. It had been more interesting in the 80486 ages. With today's hardware, all of the tactical positions are practically solved by everybody. > >Some of the positions are very interesting ( at least to me ) - as are the moves >the programs choose instead of the solution . > >For example : several programs chose Rbc1 in CCR 4 - I think it is possible to >draw some valid conclusions about the tested programs because of that . > >To make valid conclusions data is needed - that's why I appreciate everyone >testing and posting their results . > >It's only for fun - it's not like solving the problem of hunger in the world . > >One result of this testing seems to be quite clear already : the Larry Kaufmann >test can't predict the USCF rating of current chessprograms ( enough data ) . > >Does it have some other merit though ? More results => a better answer maybe. > >pete Again agreed, Uli
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.