Author: Sandro Necchi
Date: 08:07:18 07/17/05
Go up one level in this thread
On July 17, 2005 at 09:14:28, Madhavan wrote: >On July 17, 2005 at 06:49:08, Sandro Necchi wrote: > >>On July 17, 2005 at 06:26:33, Madhavan wrote: >> >>>On July 17, 2005 at 05:22:47, Sandro Necchi wrote: >>> >>>>I have been laughing a lot (maybe crying on the ignorance would have been more >>>>appropriate?)reading many wrong statements about testing and Elo lists. >>> >>>So you want only list that shows good results and huge rating difference of >>>shredder? >> >>What has all this to do with Shredder? >>Pls. be serious! >>If you do not have arguments do not say anything. >> >>>That's why I you have been laughing a lot to put the testers down. >> >>If you laugh for what you do not understand it is your problem... > >If you laugh at testers who report their results too earlier then SSDF usually >reports,then you shouldn't be laughing in the first place All tests may have a meaning and give data, but to state that SSDF list is old and replaced by new ones which do not include the same testing way is pure nonsense or lack of knowledge. Of course I refer to believe that a better way to test is better not about testing in a different way. > >" have been laughing a lot (maybe crying on the ignorance would have been more >appropriate?)reading many wrong statements about testing and Elo lists." > >were you laughing atthe testers because they don't ptoduce real data? Because of what they believe to get from those data without understanding what was missed in their way to test. I said anybody can test in a different way, but believe that they do it better than SSDF is pure nonsense. > >>> >>>>so, for those who are new and do not know, SSDF list is the best for the >>>>following reasons: >>>> >>>>1. They use 2 computers and the program complete with own book and ETG, with own >>>>gui and best setting as suggested by the programmer. >>> >>>No need if you want the customer who wants to test a program in his own comp >> >>It depends how you want to test it. That can make a difference. You must realize >>that with 2 computers the tests are more accurate. >>Isn't simple? >>If you cannot afford to have 2 computers, than it is a your problem... > >No need to have 2 computers,for those who want to play with chessmaster rather >than pitting chessmaster against some program,they would better look at the >earlier results provided to them by testers rather than wasting the time waiting >for SSDF results[for chessmaster 10000] 2 computers are needed to test chess programs between themself and not if you play against it...of course. > >>> >>>>2. They use long time controls (40/2h 20/1h; international level) only. >>> >>>people are only interested in blitz time controls,IE ask Chessmaster buyers what >>>time control they usually use.They would probably say 1-5 min for chessmaster >>>engine and infinite or 1 hour for me. >> >>Pls. use some people as most people are interested in long games and not blitz. >>If one hour for you is OK, it is fine for me. > >Talking about customers,there are many people out there who don't give more time >for engine and the same time for themselves. OK, they can do whatever they want, but not make false statements of course... I mean thay can state results at blitz or fast level and that is fine, providing they use 2 computers and the same system as SSDF. If they use one computer, one book and the same gui, to me these are false results (not correct ones). > >>>>4. They use a very wide range of programs and not only the new ones to get more >>>>reliable results. >>> >>>20 Crafty versions? >> >>PLs. be serious (again) and take a look on the extended list to know what I am >>talking about. > >There are many free programs better then crafty,for ex List,Aristarch,Yace and >Fruit OK, then one can get an indication of their strenght by testing against other chess programs in a similar way as SSDF does. If a different system is used than to me those are not reliable data. > >It still haven't been tested,I have to be serious? You have attached me personally instead of giving suggestions for testing purposes, so you got what you deserved. >serious and let them test any engine they want and let them report their result >anytime they want,IE wait for 1 year if you want reliable Fruit result from SSDF >that's pure nonsense. Maybe you do not understand English well: I said that testing is OK, but to claim that a different system is better than SSDF is nonsense. I have listed all the reasons why their system is the best and explain why. You are criticizing them only becase they do not test some program. I was talking about the testing system. They have their rules and they are indipendant, so they do what they think it is correct. If you test in the same way they do, then your data would be interesting to me. This is the point! > >>>>5. Ponder on and learning are activated. >>> >>>No need for customers,ponder on or ponder off,doesnt matter >> >>This is ESCLUSIVELY your opinion (and of a few more others). > >many customers opinion too,I might add.Most chessmaster users or even shredder >users don't about ponder mode. OK, those are wrong! > > >>>>So, anybody can test in a different way as they wish, but to claim that system >>>>is better or replacing the SSDF system is pure nonsense! >>> >>> >>>SSDF test delays to publish the result,they take tooooooooo much break.long >>>break would fit. >> >>Yes, but because they want to give reliable result. I would agree with them as >>there are TOO MANY people which are good only to make critics and nothing >>else... > >You want to wait for 5 years to see the reliable result??? Did they ever took five years? And you are telling me to be serious? >Isn't it frustrating for customers? The majority of the customers only want to know if one version is better than the previous one. You can have an idea about that after some matches and decide to buy the program or not. I cannot believe one needs to wait to find out exactly how much. Are you saying that you buy a program if it scores 5 points more or less? > >>It is not necessary to wait the new list as they are reporting in this forum the >>scores of the various matches, so one can understand if one version is better >>than a previous one after some matches. >>I do not think 5 or 6 points more or less makes a difference...so your statement >>is ridiculous. >> >>>SSDF doesn't test Fruit,still didn't test Fruit.that's irony. >> >>They will when it will be released. It seems it is strong so what is the >>difference if will be no. 2, 3, 4 or 5 in the list? >>They do not have an official book yet, so what to test? > >Fruit is released already.wake up. It is not matching what is required by SSDF...you need to wake up and a lot! Why was Shredder 8 UCI not tested by SSDF? For a similar reason. Did I criticized SSDF for that? I only said it was a pity because that was the real upgrade of Shredder 7.04 UCI, but again SSDF is indipendant and can do what they think is correct. > >Difference??? >If Chess Tiger ranks 5 and Fruit ranks 3,Its better not to buy >chesstiger,simple. Do you really think most people do this? The difference is so small that one cannot see them in their games. People are also (I would say even more) interested in playing style and available options. > >>>still didn't test chessmaster >> >>ChessMaster does not have the auto232 option, so it must be tested manually and >>that is a problem. This option is missing because the programmer does not want >>to have it included, so why to do something the programmer don't want to do? >>Pls. think or get informed before saying things... > >CM9000 has been tested. Yes, but it was HUGE work for them. I do not think they have the resouces to do it again for CM10000. To find out that it is maybe 10 points better? >so are you saying CM10000 doesn't have. > >>>still didn't test many programs >> >>They are indipendent and have rules. >> >>The good think is that they nearly always (they did an exception with Fritz >>5)apply the same rules to everybody! >> >>> >>>Testers have reported many results for the programs a day after its release. >> >>Yes, but not real data. > >yeah,you keep waiting for your so called real data.there will be more fruit >versions.Fruit will be placed #1 and wouldnt be tested. Why you do not ask them when they will test Fruit or what is needed to let them test Fruit? If Fruit will be the new no. 1 than good for Fruit programmer and motivations for other programs to get back to no. 1 > >>>For customers,SSDF list is of need to them.too much delay,unfair matching of >>>opponents. >> >>I do not agree. >>Which unfair matches if all programs are playing on the same harware, best >>setting and same time controls? > >Unfair matching of opponents I said.It means they choose engine vs engine for >themselves.excluding other well known and stonger free engines. Look they have their rules since 1980 and they keep them. They may should think to include in testing new free engines, but it is up to them to change the rules. Since these rules have been the same for everybody we cannot call "unfair matches" otherwise the programmers who did not get the same "favour" could claim "UNFAIR TREATMENT" don't you think so? > >>Agains get serious! >you should be serious. Try to lessen, get informed and learn before being ridicolous! You made silly statements and personal attack, so could not avoid to tell you. > >>>>Sandro >> >>Sandro Sandro
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.