Re: [computer-go] future KGS Computer Go Tournaments - two sections?

[Don Dailey <drd@xxxxxxx>]

> I am afraid that even this division should be completely open, ...

I meant to say "should not be completely open."   

I think most any chess programmers is going to be jaded on this
subject as there has been so much controversy in the past about this
general subject.

On ICS there used to be a huge number of "Crafty" clones, represented
by programmer wannabee's who couldn't write a single line of code.
Crafty was a very strong public domain chess program.  Many of them
ran on fast computers and fed the ego's of the wannabee's.

Then there was a real tournament (at a physical location) where there
was a Crafty derivative.  I didn't complain, but I was secretly
annoyed because I worked hard at writing my chess program from
scratch.  It didn't seem fair that someone simply starts with a
program that is already stronger than most of the participants and we
have to complete against it.  

A part of me feels the same way about any of the meta-bots or programs
like Sluggo and I feel bad for saying this because I know they worked
so hard at it.  But with all due respect everyone else starts from
scratch while they start with the  program that is already a proven
winner before they even lift their little finger.  If one of these
meta-bots compete and win how can you not say, duhhh!!!  What was the
point?

Of course on the other hand I take a keen interest in the research of
programs like Sluggo and want to shake their hand.   I would like to see 
their claims verified, not because I don't believe them but just because
that's how science is done and that's likely a big part of their motivation
to compete in tournaments.   Their claims are difficult to verify by others
which is not their fault.

While I am ranting (please forgive me), there was also a tournament
where one particular company got to have several entrants, each
different models but with the same authors.  Any mathematician can
tell you how much this increases the chances that one of their
products will win the tournament.  Of course it turns out that this
very same company was sponsoring the tournament.  What was remarkable
about this event was that one of their machines "resigned" in a won
position against one of their other machines.  There was no rule
against resigning in a won position and so it was legal.  This created
the new rule for future tournaments that you couldn't resign without
the tournament directors permission.  The rule SHOULD have been only 1
program per author.

In one of the last tournaments my program played in, I was allowed to
have 2 entrants because the tournament had an odd number of players
and one of my programs was so completely different than the others.
One was a parallel monster and the other was a simple PC program (that
actually won the 1993 international computer chess championship but
was very much outdated at this point.)  I actually had serious
reservations about this even though the tournament directors allowed
it and I detected some animosity from some of the other program teams
due to this.  I was relieved that the program did poorly and no one
cared once it lost a few games.  Secretly I realized that my chances
(as author) of winning had been unfairly increased, even if only
slightly especially in such a short tournament where anything can
happen.




> I am afraid that even this division should be completely open,
> otherwise I will enter multiple Botnoids and you will probably very
> shortly see gnu clones that are not actual derivatives, they will be
> the real thing probably represented in some cases by non-programmers
> who are smart enough to get the bot working.
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/