Premier League Predictor: Astro — Chapter 6, Exercise 2
====================================================

TASK
A fixture has three guest predictions, scored (after a result is
entered) at 40, 10, and 0 points respectively. Explain why the
prediction league table should average these point values rather
than averaging the guests' raw scorelines, and compute the correct
averaged guest figure for this fixture.

SOLUTION

Averaging raw scorelines doesn't produce a meaningful result, because
a scoreline is really two separate counting numbers (goals for the
home side, goals for the away side), and there's no rule for what
"the average of 2-1, 0-3, and 1-1" is even supposed to mean beyond
averaging each side's own goal count separately — which itself
produces a fractional goal total that was never a real, valid
scoreline anyone actually predicted. Worse, it throws away exactly
the information that matters for this table: how well each guest
actually did, not what numbers they happened to type in.

Points, by contrast, are always a single plain number representing
"how correct was this prediction" on a scale that's identical no
matter what the guest actually predicted or what fixture it was.
Averaging points genuinely answers the real question the prediction
league table exists to answer — "how well did the guest slot perform
on this fixture, considered as one comparable figure alongside user,
expert, and ai" — because 40, 10, and 0 are three instances of the
exact same kind of measurement, unlike three scorelines which aren't
comparable to each other as numbers at all.

The averaged guest figure for this fixture is:

    (40 + 10 + 0) / 3 = 50 / 3 = 16.67 (rounded to two decimal places)

This is a genuine, real number that never appears as any individual
guest's own points_awarded value, but it's exactly what the
prediction league table needs: a single, directly comparable figure
representing the guest source's overall performance on this fixture,
built the same way regardless of how many guests happened to predict
that week.

WHY THIS WORKS AS AN ANSWER
----------------------------
It explains concretely why averaging scorelines produces a
meaningless, non-comparable result, correctly identifies why points
are the right thing to average instead (a single comparable measure
of correctness, independent of what was actually predicted), and
computes the real averaged figure with the arithmetic shown rather
than just asserting the correct method.