Author: KarinsDad
Date: 16:32:41 09/02/99
Go up one level in this thread
On September 02, 1999 at 17:59:47, Thorsten Czub wrote: [snip] > >like many chess programs, cstal has a clever time control management. >it computes more about the moves that are difficult. >by giving it only 3 minutes, the clever time-management is dead. > >i have no idea what harald does. but it manipulates the result. > >>>pah - than the games are nonsense ! you do not test the programs, >>>you test something different. > >>The games are not nonsense (you like that word I guess). Neither are the tests. > >of course are the games nonsense. he is testing a broken program under >unbelievable circumstances. 1) You did not post that the program was broken until after I posted my message on this. And if this is the case, why was a broken program made available? Confusing. also he was not licenced/allowed even >to download the file. nor to publish the results. he was not, because >we wanted to stop these stupid kind of testing. > >>If you test G5 in your "prefered" configuration of permanent brains on and >>program decided time per move, it would still be a valid test of JUST THAT >>configuration (and no other configuration). It wouldn't tell you how well CSTal >>or any other program would probably due in standard game times. > >>What is it with this notion that testing MUST be done under x y z conditions and >>cannot be done under p q r conditions? Results of any testing must take into >>account the conditions set up, but that does not mean that a non-standard test >>has no validity. Granted, if the two programs had different conditions, then >>your point would be valid. But, the conditions were the same for both opponents. > >again, you don't get it: testing a fast searcher against a slow searcher giving >both 3 minutes blitz (e.g.) advantages the fast searcher. 2) This is only a theory. In fact, the exact opposite has been postulated on several occasions and there are quite a few test results that may indicate that we are finally at an NPS where the slow searchers do better at blitz that the fast searchers. 3) This thread is talking about 3 minutes PER move, not 3 minutes blitz. VERY few programs (and this includes CSTal) will change their moves OFTEN after 3 minutes. Occasionally, this may happen, but not enough to gain 100 ELO or anything along those lines (i.e. how many ELO do you think CSTal gains from it's "clever time control management?). You stated above "like many chess programs, cstal has a clever time control management". Do you think that the opponents in these games did not also have clever time control management (you implied they did) and that they were not ALSO handicapped somewhat in this type of tournament? >so you do a TEST but in fact the test-conditions to manipulate the results >even before you have started the games. 4) You ignored what I wrote. I SAID several times that you CANNOT take those results and interpret ANY other type of information from them such as chess program strength in standard tournament conditions. This has nothing to do with mainpulating the results before starting the test. It has to do with being aware of what results you can make BASED on the test conditions set up. 5) These tests of Harald's do not prove ANYTHING about chess program strength. You and I are on agreement on that. However, that does not mean that you cannot glean information from this type of test. Think about it. KarinsDad :)
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.