Re: [computer-go] Protocol B

[David G Doshay <ddoshay@xxxxxxx>]

On 28, Jul 2005, at 8:34 AM, Ben Shoemaker wrote:

> Does the protocol enable a program to "cheat" and force another 
> program to
> lose on time?  I don't think so.  If the cheating program is playing
> worthless moves, the refutations should be obvious and not require a
> significant amount of time to play.
Unfortunately, this is not the case. When SlugGo played in the Cotsen 
Open
one opponent was loosing badly on the board, by more than 100 points. He
looked ready to resign when I casually commented that he was not going 
to
win on the board, but might win on time. He was able to find moves that 
were
not going to result in a win, either from an equally strong human or 
from
SlugGo, but he did succeed in running SlugGo's clock down and he won
the game. The distinction is that "worthless" can still be tactically 
complicated.

As far as I was concerned this was a fine result: He got what he wanted,
which was a win in the tournament, and I got what I wanted, which was to
see that SlugGo could beat a human 8 kyu if the clock had not been an 
issue.
Pareto optimality.

Cheers,
David



_______________________________________________
computer-go mailing list
computer-go@xxxxxxxxxxxxxxxxx
http://www.computer-go.org/mailman/listinfo/computer-go/