Author: Uri Blass
Date: 15:46:27 10/07/01
Go up one level in this thread
On October 07, 2001 at 18:10:07, Francesco Di Tolla wrote: >On October 05, 2001 at 20:52:36, Christophe Theron wrote: > >>You could say the same for every additional ply depth: a ply is a ply, a piece >>of knowledge is a piece of knowledge. > >An so what? > >The point is that the piece of extra knowledge remains there, the extra plys are >reduced! That's the whole point. > > >>I believe there is a strict equivalence between search and knowledge. They are >>the same thing. Like energy and matter are the same thing in modern physics. > >The comparison has nothing to do (you're talking to a theoretichal physicist). > >>So if there is dimishing returns from deeper searches, there is dimishing >>returns from better knowledge. > >I don't see way. >I see clearly that the difference in depth between two programs reduces >increasing the speed (due to the nonlinear growth of the number of nodes). > >On the other side I see no reason for what you imply: that understanding coming >from a better evaluation fades as you increase the total search time. I can give you a simple example that knowledge gives nothing if you search deep enough(I do not believe that it is the general case but I can understand Christophe's point). Imagine the programs do not use tablebases Is it important to know how to win KQ vs K At small depthes it is important because a program without this knowledge may draw these positions. At big depthes it is not important because the program without this knowledge is going to find the win by search. The only important knowledge at big depth is to know that KQ wins againt a lone king and the computer can safely give the same evaluation for all KQ vs K positions. Uri
This page took 0.01 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.