I am thinking that, within the circuit's
universe, talent/ability/scoring is probably closer to that
of the "pyramid" model.
Talent and ability
in quizbowl might be normally distributed across a
general population. (I think this is as good an
assumption as any for a question that we can't really answer
anyway.) Except that people on the low side of the median
have almost certainly found something else to do with
their Saturday afternoons.
So what you have
left is the right half of a bell curve (and probably
less than that - in just about all cases, even the
least-active buzzer rock on the lowest-scoring team is
probably well above the 50th percentile among those
eligible for these competitions.) The only evidence I have
of this is anecdotal : even low performing circuit
teams regular rack up points at CBI RCT vs. once-a-year
competition, and even that competition had to get there with
top performance at their school's IMs.
Someone at some point could sit down and figure out the
typical PPG (or some other easily obtainable metric)
distribution to figure out what it looks like. Maybe it does
look like a bell curve with the bottom half lopped off
; maybe it doesn't.