RE: computer-go: prevent cheating

[Sai To Wang <swang@xxxxxxxxxx>]

I had a refinement on the work around for timing problem.

What about in the replay mode, the server donīt provide any
information other then that they are in replay mode, and the
Go-playing program send the decision move of every search-depth
level (256) then the server take the move which indicated by
the search-depth level in the game record.

So both program had zero knowlage other then itīs in replay mode.

Of cause the Go-playing need to be run in a iqual or better machine
and all the move need to be timed and calculate the proximate time
to find out that was the same program can finish the same game with 
the machine specified by the player.

Please also reply to [stwang@xxxxxxxxxxxxxxxxx] thanks.



> -----Mensaje original-----
> De: Mika Kojo [mailto:mkojo@xxxxxxxxxxxxxxxxx]
> Enviado el: martes 19 de octubre de 1999 18:18
> Para: Sai To Wang
> Asunto: Re: computer-go: prevent cheating
> 
> Is this equivalent to the following game protocol;
> 
>  1. Player A thinks about a move. 
>  2. Player A sends the move and the thinking info to the referee. 
>  3. Player B thinks about a move. 
>  4. Player B sends the move and the thinking info to the referee.  
>  5. goto 1
> 
> Now clearly this protocol is not secure against active cheating. The
> cheater could put anything it wants to the "thinking info". Even the 
> made move. 
> 
> However, if the information is send to the referee _before_ the move
> then it could be secure. In this case though, I cannot see how to help
> the timing. 
> 
> Of course, I might have understood this incorrectly. 
> 
> 
> -- Mika
>