Author: Christophe Theron
Date: 21:40:40 10/17/00
Go up one level in this thread
On October 17, 2000 at 17:31:24, Ratko V Tomic wrote:
>Thanks Christophe. I had always found it amusing when the "old guard"
>tries to make judgments on where we are and what is next. At the end
>of 19th century, the "old guard" proclaimed that physics was just about
>done, some extra digit left to calculate here and there, and couple
>small clouds to clear. Or, in early 1950s, the IBM proclaiming that
>world may need as many as 10 computers. Or, a bit later, IBM turning
>down the inventor of Xerox machine, wondering why would anyone need
>to make paper copies. And there are thousands of other such instances,
>who needs airplanes, who needs phones, who needs cars, who needs mini
>computers, who needs micros, who could possibly need more than 640K
>of RAM, etc.
>
>Bob's categoric proclamations often give me daja vue of those wise
>folks, whose experience and knowledge tie their imagination to
>the ground and they can't see any more what is possible beyond their
>simple little models. If your program evalutes position as +3 when his
>program counts wood as equal, for him this automatically means that,
>should one remove your piece, you will evaluate position as equal,
>even though it is then clearly lost. And that's the end of the story,
>and there can't be any other way.
Yes, and as you clearly pointed out, this is wrong.
There is nothing wrong with evaluating an attack at the value of a rook or even
a queen.
It's just a question of efficiency. Does it work or not?
Many programs would evaluate a very strong attack to be equal to the value of a
pawn. Because when the author tried to evaluate it higher, his program began to
play worse and worse.
This is just demonstrating how BAD these programs are at evaluating attacks. It
does not show that an attack cannot have a value higher than the value of a
pawn. Any chess player, even the weakest, knows that!!!
So, being unable to teach their programs what the game is all about, instead of
writting chess programs they write checkers programs.
These programs don't play chess, where the goal is to capture the opponent's
king, they are playing checkers, where the goal is to capture all the opponent's
pieces!
I must add immediately that I have been guilty of this myself. Until recently I
was writting checkers programs, and sold them as chess programs. :)
But my latest version is a REAL chess program.
Christophe
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.