Author: Mogens Larsen
Date: 03:03:52 05/22/02
Go up one level in this thread
On May 21, 2002 at 13:34:32, Dann Corbit wrote: >What about a novelty discovered from long time analysis by some program? If the >GM remembered the novelty and used it, wouldn't that be learning from a >computer? Without daring to presume to know how a Grandmaster prepares, my impression is that it's mostly the other way around. That is, checking variations of manual analysis. In general I think that letting a program contemplate a certain position is less efficient than verifying lines analysed without a program. Both for blundercheck and for finding refutations and defense against "missing" moves. >What about probing a desired line for deep tactical shots? That isn't efficient either IMHO. If I were to be a pretend GM again, it would seem more relevant to analyse a broad variety of responses to, say, a novelty. Sufficiently deep to imply the plan to follow afterwards. Having a couple of deeply analysed lines is more susceptible to being avoided. And very time consuming as well. >I don't think computers will show GM's any techniques. But I think they can >demonstrate facts. The GM's can make important inferences from these facts. Only if the facts are general rules. Knowing the correct order of moves in a complicated five piece endgame is handy. But if the method can't be applied sensibly to other positions with the same pieces, then it is essentially worthless. Maybe there's a couple of unknown general rules hidden in the EGTB. I just can't remember hearing of any. Regards, Mogens
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.