Computer Chess Club Archives


Search

Terms

Messages

Subject: Learning Functions and "Fair" Competition

Author: Fernando Villegas

Date: 07:39:23 11/23/97


I confess that I have been reading with some amazement the great amount
of post published here, in RGCC or elsewhere, in which several ideas are
proposed to get a fair competition between programs on the ground of
eliminating the so called “learning function”. So, some people talk of
games performed without the book, or begun from a middle game position
and of course of tournaments  where doubles are forbidden.
I don’t understand this line of reasoning such as I did not understand
the old discussion about how to eliminate the hardware power of Deep
Blue in order to get a “fair” competition with PC programs. Such as the
hardware of Deep Blue is a constitutive part of his Being and so
disabling it has no sense at all, I understand that with even more
reason any learning function should be considered a constitutive part of
a chess program and to think unfair his utilization is a mighty absurd.
Is not memory and learning an special aspect of intelligence in any kind
of beings? Can you think of intelligence without learning? Why a
learning device should be more exogenous to a chess engine that any
other specific device? How about extensions? Or of any device to get
faster search? It could ‘nt be said that is  “cheating” to win a game
“just because” the program search a lot more nod than any other?
Whats’ more, the learning function has became something as a known
technique that can be used by anybody in the field. Of course, those who
does not use it are in disadvantage, but then they would be in
disadvantage also if by any reason they were not incline to use, let us
say, recapture extensions. In fact, if the argument of learning function
as a kind of unfair device and almost cheating device is valid, then the
programmer of any low level program  could say that for competing in
fair conditions against, say, Rebel 9, Ed Schroeder should then disable
all the functions that gives such a tremendous power to his program.  I
don’t understand this and so, If I am mistaken, I wait with anxiety a
revelation.



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.