That seems like the most likely way of dividing it, to me. I'm not entirely comfortable with endorsing that model (I don't know enough about it, basically), but it looks plausible at least. (I'm not entirely sure we're going to divide it up at all, at this point, but the code's written in such a way as to at least allow it.)

I suppose it'd be interesting to allow the different M-B groups to optimize the quotes that the groups get separately, rather than having all the votes go into the same database - I don't think it'd be all that hard to set that up, and it'd give us some interesting data on M-B.