Re: [computer-go] KGS Tournaments: Cheating

[Mark Boon <tesujisoftware@xxxxxxxxx>]

On Mon, 7 Mar 2005 13:25:04 +0000, Nick Wedd <nick@xxxxxxxxxxxxxxxxx> wrote:
> The easiest way to cheat in a server-held Computer Go event is to enter
> a program which is able to receive "hints" from its operator.
> 
> A current proposal for avoiding at least this form of cheating is as
> follows.  Each participating program is sent, as an executable, to a
> "Trusted Person" or TP, who has an appropriate platform to run it on,
> and will run the copy that enters the tournament.  A TP must be trusted
> by the programmer not to steal his code, and must be trusted by KGS and
> the sponsor (and preferably by the other participants) not to cheat.
> Different programmers may trust different TPs.  The TP will receive an
> internet-transaction logger from KGS.  He will close all other
> applications that access the internet, and run the logger, kgsGtp, and
> the program.  At the end of the event, he will sent the log to KGS for
> examination.
> 
> Another proposal is that programs must play reproducibly.  Then, if
> there is an accusation of cheating, the accused program can be asked to
> reproduce the sequence of moves.  However, some programs don't play
> reproducibly, and can't be forced to.
> 

I don't think the TP solution will find many followers. And it also
introduces too much extra bureaucracy.

Maybe requiring the program to play reproducibly is reasonable. Some
programs use a random-generator in order to obtain more variety of
play. But in those cases it should be possible to generate the same
moves by making a simple feature where the same seed can be used to
reproduce the same numbers. Still its not fool-proof.As in sports, you
can have 'doping tests', where any program that wins a prize has to
reproduce one or more games. But how would you prevent a cheater from
sending the same hints again? Or worse, reproduce the moves from a
saved file, cleverly sent with a similar interval as in the original
game.

Of course there could still be programs that can't reproduce play for
other reasons than a random generator. One source could be bugs.
Forcing moves to be reproducible may actually be doing a favour to
those programs :) Another source for irreproducible moves is
time-control. If a program adjusts the amount it reads based on real
time used, you may occasionally get different moves when trying to
reproduce the game. This is be a little harder to crack. You may end
up requiring programs to build in a mode that will guarantee
reproducible game just for online play.

A real, cheat-proof system for online tournaments has not been found yet AFAIK.
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/