Author: Komputer Korner
Date: 11:32:42 05/14/05
Go up one level in this thread
On May 13, 2005 at 20:07:45, David Rudel wrote: >In lay terms, backsolving evaluates a position assuming best play from both >sides. Of course you have to be correct about the evaluations of the ending >positions, and you have to assume that you have all viable moves, but even with >those two huge assumptions, backsolving is a better method than simply looking >at which move has historically done best. > >One reason that looking at historical statistics is that when a new good move is >found, it may be quite succesful until it is refuted, but then once it is >refuted, the whole line may be abandoned. Thus, from a statistical point of >view the line looks great [because the statistics will be skewed by all the >earlier games where the line was doing well], but trot that out on the board and >you may be well be playing a busted line. I never said anything about move statistics. They are notoriously inaccurate because of refutation novelties. Backsolving doesn't evaluate a position. It merely attaches a symbol(mathematical number or otherwise) based on subsequent symbols. See a separate post for the case where Backsolving would be useful.
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.