[computer-go] A new pairing system idea for CGOS

[Don Dailey]

Yes, simulations are good.   N may need to change with the number of
active players at scheduling time.  

I generally avoid pairing the same 2 consecutively, although there is no
particular need to do this.

If you do the sim, try top down,  and random (choose a random player,
then a closest of N opponent.)  

CGOS currently uses this basic algorithm, using the random version of
it,  but I don't schedule discreet rounds so it's not so simple.  This
is also a big part of the problem.   (Actually, I'm not convinced there
is a big problem, but there is at least a small problem with opponents
who play too fast and monopolize the scheduling.)  

- Don





On Thu, 2006-10-05 at 14:58 -0700, Christoph Birk wrote:
> On Thu, 5 Oct 2006, David Doshay wrote:
> > I fear that too much of a "top down" system will not have
> > enough mixing between levels ... but I also fear that I am
> > being too much of a critic with too few good ideas for the
> > solution.
> 
> That's why 'N' should not be too big ... I will do
> some simulations.
> 
> Christoph
> 
> _______________________________________________
> computer-go mailing list
> computer-go at computer-go.org
> http://www.computer-go.org/mailman/listinfo/computer-go/