Author: Tony Hedlund
Date: 09:53:52 02/18/03
Go up one level in this thread
On February 17, 2003 at 06:29:23, Rolf Tueschen wrote: >On February 16, 2003 at 13:21:39, Tony Hedlund wrote: > >>On February 15, 2003 at 07:12:10, Rolf Tueschen wrote: >> >>>On February 15, 2003 at 05:24:43, Tony Hedlund wrote: >>> >>>>On February 14, 2003 at 16:27:31, Rolf Tueschen wrote: >>>> >>>>>On February 14, 2003 at 13:32:16, Tony Hedlund wrote: >>>>> >>>>>>On February 14, 2003 at 09:27:26, Rolf Tueschen wrote: >>>>>> >>>>>>>On February 14, 2003 at 08:43:12, Bob Durrett wrote: >>>>>>> >>>>>>>> >>>>>>>>Excellent points. The "bottom line" is that SSDF presented their findings >>>>>>>>properly, but the problem is in interpretation. SSDF cannot be held responsible >>>>>>>>for errors in interpretation. >>>>>>>> >>>>>>>>Bob D. >>>>>>> >>>>>>> >>>>>>>Wrong conclusion. I tried to explain the points but apparently it's a bit too >>>>>>>difficult. In short : If you use a system of statistics you are not allowed to >>>>>>>make your own presentation. The presentation by SSDF is FALSE. That is the >>>>>>>point. False and unallowed. Instead of 1., 2., 3., they should say 1.-3., not >>>>>>>should, but must, if the differences in the actual results are way smaller than >>>>>>>the error in the tests itself. Is that impossible to understand? >>>>>>> >>>>>>>Rolf Tueschen >>>>>> >>>>>>Then the right presentation is: >>>>>> >>>>>>1-10 Shredder 7 2801-2737 >>>>>>1-10 Deep Fritz 7 2789-2732 >>>>>>1-11 Fritz 7 2770-2711 >>>>>>1-2? Shredder 7 UCI 2761-2638 >>>>>>1-15 Chess Tiger 15 2753-2700 >>>>>>1-15 Shredder 6 Pad UCI 2750-2703 >>>>>>1-16 Shredder 6 2750-2689 >>>>>>1-19 Chess Tiger 14 2744-2684 >>>>>>1-19 Deep Fritz 2741-2680 >>>>>>1-19 Gambit Tiger 2 2739-2681 >>>>>>3-2? Junior 7 2715-2659 >>>>>>4-2? Hiarcs 8 2707-2657 >>>>>> >>>>>>and so on. >>>>>> >>>>>>Tony >>>>> >>>>>Thanks for the fine joke, Tony. Perhaps you lay your figer into the wound! >>>>>You want to have a number one, right? Then you make tests, just like you do, >>>>>fair and correct. And then you come into the period where you must evaluate your >>>>>results. You see that you have no clear umber one. Now two possibilities: >>>>> >>>>>1) You go on into decisive mode and do further tests, the "list" date can wait. >>>>> >>>>>2) You stay to your traditions and show up with your list. But then, please, do >>>>>NOT present the list either in the classical way, nor in your joking Mr. Bean >>>>>version, but simply make such packages: >>>>> >>>>>1.-3. A B C >>>>>4.-5. D E >>>>>6. F >>>>>7.-10. G H I >>>>>etc. >>>>> >>>>>Tell me please, where the problem is with this method? >>>> >>>>Why just three strongest engines? With the margin of errors Gambit Tiger 2 could >>>>be as strong as the other top engines. I find Mr. Bean's version more logic then >>>>yours. Could you please explain your method further. >>> >>> >>>SSDF has good statistics experts. Consult these experts and you will understand >>>why Gambit Tiger 2 could NOT be number one. My first three was a pool where all >>>could be number one. Only Shredder 7 UCI could be included, but my example was >>>more a demonstration of such a list. It's not MY method. It's simply what >>>careful researchers would do if they had your results. Perhaps you don't know >>>it, Tony, but the presentation of the results must have a base in the results. >> >>What do you propose SSDF do exactly? Give me a clear example of how you would >>present the data. Don't give me this A, B and C. You have the result, wich >>programs are A, B and C? >> >>>In other words it might well be that one day you will have a clear number one. >> >>The bottom line is that when we reach a margin of error close to zero, then we >>can claim a number one? When will that happen? After 10 000 games by each >>entrance? >> >>>Or do you believe that your method guarantees the eternal status quo? >>> >>> >>> >>>> >>>>>Is it because you have >>>>>kind of strong wish to present a umber one by all means? >>>> >>>>Do you also think that FIDE shouldn't have a number one on there list? Is >>>>Kasparov really the best player? >>> >>>Please do not seek for outside help, when you run out of arguments in favor of >>>your own presentation. >> >>FIDE, ICCF and SSDF all have a ratinglist. And we all use professor Arpad Elo's >>metod of measure strenght in chess. And yes I argue for our way of presentation. >>ICCF's number one Ulf Andersson have played 25 games! Figure the margin of error >>there. They probably don't have any careful researchers. >> >>>> >>>>>Please let's simply >>>>>discuss this little topic. If you tell me, listen, Rolf, I am not allowed to >>>>>tell you, but you are right, that a umber one prog is very important for us. >>>> >>>>It seem to be more important to others. >>> >>>Yes, that was my deeper assumption. Could you give more details? >> >>Details? >>People here at CCC seem to be looking forward for our next list, to see wich is >>number one. And then they congratulate the programmer. And of course the >>commercials use it in there advertisement. As they always has. When we started >>our list, it was as a complement to our reviews for new programmes. >>Personally I'm not interested in wich program is number one. I'm more interested >>in how the different engines are playing. > >I can well imagine your personal sentiments and I have great respect for your >efforts with SSDF as a whole but you can't stop history's progress. When you >played move by move with the ancient chessboards your dedication and hard work >was really sensational and people got results for their virgin background. Today >- with autoplayed games - you have more time to do sound statistics. However, if >simply the top programs do not differ that much then you can't call out a number >one. Or you play millions of games. But who guarantees you that then you will >have a clear first? No - you should accept the actual reality. And that is >equality among the top entries. That's why we have the margins of error. So the intelligent users can make that interpretation. >You are misleaden if you think that the thankfullness of the CC users was linked >with your presentation of a number one. It was because of your general efforts >to the best of CC. And the business world at that time was very coloured. But >today we have a single important company. Do you want to do your job for them >and their marketing interests or for the users around the world? You must >accept that if statistically you have no clear first then you can't present a >number one program. What does that bother you??? You are independent! But >independent does not mean naive.Why don't you consider the consequences of such >strange events: Fritz8 is out for months and you don't test it. I read that you >wait until ChessBase will send you a copy. But that then would no longer speak >for your independent tests. We also wait for a new version of Yace and some copies of CM9000. >Because factor time of testbeginning always was a >factor. All such dangers and difficulties you could avoid with sound statistics We already have sound statistics. It's your OPINION that we don't. >and certain basic guidelines. You must become independent of such marketing >decisions by ChessBase. Yes we depend on getting free copies of prgrams since we dont have the economy to buy copies to all our testers. >Don't ask me for the details. I am not a member and I was defamated long enough >by your collegues in the staff. > >Rolf Tueschen > > > > > > > >> >>>Rolf Tueschen >>>> >>>>>Then, Tony, I am out of the debate, because I had great respect for your amateur >>>>>approach. Comps are not cheap either. etc. To make it clear. I would not oppose >>>>>sponsering. But if you said, but Rolf, look, we have a real number one! That is >>>>>the exact result of our statistics. - Then however, I will continue to ask >>>>>polite questions. >>>> >> >>The exact result of our statistics is the way Mr. Bean interpret the list. >>You choosed not to comment on this, why? >> >>Tony >> >>>> >>>>>Rolf Tueschen
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.