If you're unhappy with the way players are ability-grouped, you should use random grouping over an entire season. This is what the Pittsburgh and New England leagues do, and it's worked out pretty well.
I'm a big fan of actually playing head-to-head against other opponents, with wins and losses against those opponents mattering, instead of against a collective "everyone" as Kidforce does. But that's just my opinion.
The other way to go would just be to have whoever is in Group #1 at the end of the season compete as the A finalists.
I don't know enough about FSPA scoring, but I do know that it punishes very poor weeks of play above all else -- don't score a 3 in Week 5 next time, Jeff