Author: Komputer Korner
Date: 00:22:34 12/21/98
Go up one level in this thread
On December 20, 1998 at 16:49:08, Christopher R. Dorr wrote: >First off, I eliminated the multiple include levels in order to make this >clearer. There are several points I wish to make. > >1. KK, your catagorizing me as 'emotional' does nothing to advance this >discussion. Futhers, as Reynolds Takata rightly pointed out, you seem to be the >one deeply invested in making sure that everyone agrees with your assessment of >CM6K as a 'Toy' program; you started the thread, you chose the inflammatory >words. > >2. So far, you (someone who may not currently be an expert-level player, but who >has 'maxed out' at 2023) has said that he knows better than at least 5 titled >Masters what a chess program should do and what it shouldn't. And that he knows >how to train better then me (USCF Life Master), Reynolds Takata (USCF Life >Master; FIDE Master), and Gregory kaidanov (USCF 2650+, FIDE 2600+, Grand >Master). This is pretty absurd to anyone who looks at it. > >3. You answered none of my requests for your qualifications other that posting a >perhaps out-of date rating. If you are going to make these kind of statements, >you probably should have some credentials for doing so. All I am asking is that >you post them, so that everyone has some background upon which to judge your >'expertise'. I'm not at all saying that we shouldn't pay attention to you >because you're not a master, but that someone reading this thread might want to >decide for himself whether to listen to Me, and Reynolds, and several other >Masters, or you. In order to do this, he needs more information. Why won't you >provide this? > > >4. A PV and an 5 move analysis on screen (what you are proposing as 'the best' >computer based training program) simply fails on a number of points > >a. the computer has no way of expressing a plan in a PV. > >b. The computer has no way of saying why a given line is better than >another...the student is simply left with 'Bb5 is +1.25, and h3 is +1.35. This >information is often entirely meaningless. > >c. The differences between evaluations of less than a pawn or so are often >entirely meaningless for most players. Any program is going to use an evaluation >function with features not understandable to non-programmers. Why is Bb5 +1.23, >and Why is h3 +1.35? The PV's are only going 5 moves deep. If the evaluation is >this close, it's probably not a material disparity; it's an artifact of the >evaluation function. This artifact give *no* information to the student >whatsoever. > >d. The student learns nothing about *how* to judge a position. It's great that >Fritz thinks the position is +.34 for white...how does the student make a >similar decision, and how does he use this? > >e. The student is encouraged to judge a given position as a single and >disconnected position. Each time the comp sees a new position, it starts it's >evaaluation anew (forgetting about hash tables for the moment. This may be how a >computer plays, but it is *EXACTLY* how a human *SHOULDN'T* play; the student is >learning bad habits, and is disconnecting the position from the flow of the >game. > >Your ideas about why CM6K is a 'toy' program are related only to this and the >opening book editor. correct? Well I think it's pretty clear that a *large >number* of strong, serious chess players disagree with you on this account. > >Wh can't you simply submit to 'well, we have a difference of opinion' and let it >go? You keep saying things like "Kaidanov obviously doesn't use a chess program >for training" which, BTW he does, and "You guys are simply wrong..it is 'the >single most important feature of a chess program". You seem to have to prove >yourself right, and Me and Reynolds Takata, and the other Masters who have >commented on this wrong. First off, it won't happen...qualifications *do* make a >difference, and second, I'm, confused as to why this is so important to you. > >Chris Dorr >USCF Life Master You are being emotional about this because you have no argument. You say that looking at the PVs will not enable one to get a plan from the position. So just how do you propose having the program indicate a plan for the user? The fact is you can't, so why would you bring this up as your main reason why you propose that you don't agree with me? My simple point is that CM6000 makes it much more painful to do player player analysis with the engine PVs showing. One has to type extra keyboard shortcuts with every change of direction and move. Other top programs make it easy. This ability is an extremely effective way to get to the truth of a position. How else would you propose to do it? I am not talking about listening to tutorials for beginners and those that need instruction. I am talking about how you study chess. Assuming that Kaidanov is not looking over your shoulder, then how do you go over your tournament games. Think of what you did before computers were here. You looked at each position and took back moves and tried out certain others and you moved back and forward through the game to see where you went wrong. Every chess player does this. So just because computers come along why would you change the basic way to study your games? If you are saying that because computers have come along you now simply set the auto annotate to overnight analysis and look at it in the morning, then you are sadly mistaken about the value of this. It has some merit of course but nowhere near the merit of the usual back and forth analysis method. My point is that the back and forward analysis method is enhanced by the computer programs; not eliminated. If you don't do back and forward analysis, then I don't know how you got to master level. You must have had Kaidanov looking over your shoulder the whole way. Not everyone has that luxury. -- Komputer Korner
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.