Computer Chess Club Archives




Subject: morph(C) -- can this really work ?

Author: Georg v. Zimmermann

Date: 01:25:44 10/13/99

Go up one level in this thread

about morph :

how on earth can a 1ply (see morph-webpage) ever be enough to play decent chess?
Even a beginner searches maybe 2 plys.

isn't it a huge problem that when morph lost how should it know _why_ it lost. I
mean if in the first game it lost q. vs. n. on move 10 and it doesn't know about
piece value, and gets mated on move 80 : how should it know that exchanging
queen for knight is wrong ?

On the other hand, this is a very interesting approach, because I believe that
"you can't teach what you don't know" = computers will never cross a certain
playing-strength (excluding faster hardware developements) because _we_ don't
know whether a bishop is worth 2.9, 3.0 or 3.2 pawns.
But maybe a program can learn that ...


This page took 0.02 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.