Author: Drexel,Michael
Date: 17:09:35 09/29/03
Go up one level in this thread
On September 29, 2003 at 18:50:07, Jeroen van Dorp wrote: >On September 29, 2003 at 16:22:06, Drexel,Michael wrote: > > >>What are you talking about? > >You are writing your question directly under the answer. I repeat it for your >convenience. You cannot prove that an engine finding the right move for the >wrong reason is the strongest engine. Strongest? Prove? That was never my intention. I just did some automated tests to get an idea myself and posted the results. Everyone can decide for himself what to do with that. > >You stated yourself that it's quite possible that an engine changes the PV and >evaluation after a longer time. Anyone with experience with test results knows >this is not a "statistical" or "hypothetical" question, but a very valid >situation that often occurs. > >That's why you were given the advice to repeat the suite with the full time, and >see it for yourself. Why should I? I am convinced that my results are sufficient to get some insight. As expected I got exactly the same results with extra ply 99 at WAC for 2 engines. 288/300 0.09s/0.13s for Chessmaster Pillen and 279/300 0.38s/0.42s for Shredder 7.04 Not surprising of course (1 sec maximal solution time). > > >>In most of the cases an engine finds the right move for the right reason. >>This is just a question of probability. >>There are cases when it finds the right move for the wrong reason, but that is >>true for all the other programs as well. > >And so how do they compare to each other? After all that is what you were trying >to assess. You were making a comparison between engines, and chose one engine >from the batch: > ><quote> >Q: Why? >A: to get an idea which programs are best suitable for interactive Analysis. ></quote> > >The method you use is not a proper method. If that remark bothers you, it >doesn't mean the people mentioning it are wrong and you are right. It simply >bothers you. In that case it would be better not to make the claims you have >made, to avoid irritation. Show me where I made any claims that my method is the right method. You are the person who makes claims and they are not justified. Just read carefully. I would like to get an _idea_, not a final verdict. IMO my method is sufficient to get an idea. > > >>How often do I have to repeat it? >>It is not important for me whether the PV is wrong or not or whether it discards >>the move or not later on and never returns again. > >There's a bell ringing somewhere that you wanted "to get an idea which programs >are best suitable for interactive Analysis". If the PV is wrong the analysis is >wrong. > >Now you claim that it's not relevant if the PV is wrong or right. Yet you claim >to test for finding the best engine determining the right PV in the shortest >time. > Huh? Where did I claim that? > > >>Yes >>Eventually I do exactly the same during interactive Analysis. > >Then your interactive analysis is wrong also. But that doesn't matter. We're >talking about methods here, not your notion of practical useability. >If you feel "some move and PV returned from the engine with some evaluation" is >sufficient for you, that is _no_ problem. It's _your_ choice. >However the method you use constinues to stay not the right one for answering >your original question. > > >>It is my task to find the right move with the correct assessment. > >You're welcome. As long as you're aware that you started this thread with a >statement about finding out which _engine_ was the best for that, and with the >results you presented you stated that the _test suite_ gave you the answer to >your question. No. I just posted a result. Nothing more or less. I even didnt post my own conclusions. It was not about your task finding a move with the correct >assessment with your own brain. > >If you have found the engine that caters you, that's okay. It doesn't mean your >test method is valid. > >If that bothers you, simply don't publish the test results. > What bothers me is, that some people tell me that my method is wrong and their method is right. This is utter nonsense. Michael > > >J.
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.