Author: James Robertson
Date: 07:30:20 02/09/99
Go up one level in this thread
On February 09, 1999 at 09:47:26, Dann Corbit wrote: >If, after location and execution of a 'best move' the computer's evaluation >function has a very large negative value (and possibly still dropping) can it >really be said to have *found* the correct move? > >It seems to me that we have two possibilities: >1. The program is on the right track, but has not really understood *why* the >move is good yet. In other words, "we have an accidental solution." >2. The move is actually not helping matters. You are going to lose anyway. >The test suite overlooked a good response by the opponent. > >In either case, I don't think a true best move was found. Shouldn't a best >move >promise at least a draw and not a loss? Opinions? I suppose it depends on what the test suite designer was thinking.... I.e. if he writes a test suite where you have to find the move to delay mate as long a possible then (hopefully) every position will result in the side to move being mated. James
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.