Re: [computer-go] KGS Tournaments: Cheating

[Stuart Yeates <syeates@xxxxxxxxx>]

On Mon, 7 Mar 2005 13:25:04 +0000, Nick Wedd <nick@xxxxxxxxxxxxxxxxx> wrote:

> 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.

Surely this merely ensures that the cheater cheats reproducibly too?
Unless the software is mass-market or open source and anyone can
reproduce the game, move by move reproducibly doesn't seem to gain you
much. Even then you'd need to restrict program authors to playing
released versions, rather than the bleeding edge experimental
versions.

There are some ways of making cheating harder, including:

* Playing a game at a random time over a 24 hour period or playing
games continuously for 24 hours, working on the assumption that the
program author is not going to be watching it continuously. This works
to the disadvantage of authors who don't have robust clients which may
crash during the period.

* Having an independent person (i.e. a local dan level player)
"witnessing" the playing of the game at each end.

* Have the game played in a remote VM (i.e. VMWare), in which all I/O
can be monitored, or cut off, during game play. This is a technically
challenging solution.

None of these are perfect. 
 
cheers
stuart
_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/