Author: Torstein Hall
Date: 14:45:03 05/26/02
Go up one level in this thread
On May 26, 2002 at 11:19:19, Martin Schubert wrote: >On May 26, 2002 at 11:04:01, Torstein Hall wrote: > >>On May 26, 2002 at 11:01:04, Martin Schubert wrote: >> >>>On May 26, 2002 at 10:42:34, Torstein Hall wrote: >>> >>>>On May 26, 2002 at 07:49:26, Rolf Tueschen wrote: >>>> >>>>>On May 26, 2002 at 05:09:33, Torstein Hall wrote: >>>>> >>>>>>>Another point: if you took a look at the list where Shredder was leading you >>>>>>>could see that the leading programs had played their games against totally >>>>>>>different opponents. So you can't compare the ratings at all. >>>>>> >>>>>>If you can not do that then I think you can forget about rating. I'm playing >>>>>>different players based on rating and of course often we have not played the >>>>>>same persons. That is one of the reasons we have rating! >>>>> >>>>>This is absurd. I assist Martin Schubert that _testing_ could not allow >>>>>deliberately chosen opponents. We are talking about rankings in test series, >>>>>_not_ in real life tournaments. >>>>> >>>>> >>>>>> >>>>>>>My suggestion: the top programms should play the same opponents to make it >>>>>>>possible to compare their results. >>>>>>>If I remember right it happens quite often that a program is very strong in the >>>>>>>first rating list it appears in (where it plays against weak opponents). In the >>>>>>>next rating list where it has to fight the tough ones it falls back in the >>>>>>>rating list. >>>>>> >>>>>>That is what the error margins are for. I think the rating normally stays within >>>>>>this limits. So for a given program that has got a SSDF rating of say 2600 +/- >>>>>>43 You can say with 95% (if I remember right) confidence that the program has a >>>>>>rating within the range 2557 - 2643 >>>>> >>>>>This is absolutely false. THe error margins have _nothing_ in principal to do >>>>>with different opponents (on different hardware actually)! The margins are >>>>>simply a consequence of the statistical maths. >>>>> >>>>>Rolf Tueschen >>>>Who are you arguing with? >>>> >>>>The absurd thins is that I never has sayed what you say is absurd!!!! I was just >>>>reading what the numbers meen! And that is we can tell a rating with 95% >>>>confidence inside this margins! >>>> >>>>But another thing Martin did say was that we can not use the numbers when we >>>>have played different players. I disaggree strongly to that, as long as we are >>>>talking about the same pool of players. If it was not for that, rating numbers >>>>would be utterly useless. (And maybee they are....... :-D ) >>>> >>>>Torstein >>> >>>Not the ratings are useless. Trying to do statistics with ratings is useless >>>under these circumstances is useless. >> >>What makes the statistics useless? Book learning? And is not that part of the >>program? Just as the opening book? >> >>Torstein > >For doing statistics you need independent experiments. With book learning this >is not the case of course. >If you think that book learning is part of the program or not doesn't have >anything to do with statistics. And if you take a look at any statistical test >or something you usually got the condition "independent indentically >distributed". So if you don't have iid Variables you can't do this test. And if >you do it the results are worthless. Well..... We do statistics on peoples opinions, like opinion polls. And who sayd that people where independent of each other? They talk and iterfere with each other in many ways....... To be serious, I do not think the book learning interfere that much. It just stops the program to repeat stupid lines it do not understand. They still have to play the game when they come out of book. Torstein >> >>> >>>> >>>> >>>> >>>>> >>>>> >>>>>> >>>>>>Torstein >>>>>>> >>>>>>>Regards, Martin
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.