Premier League Predictor: Astro — Chapter 1, Exercise 2
====================================================

TASK
Explain why this app averages multiple guest predictors into a single
"guest" figure for weeks with more than one guest, rather than
tracking each guest as a permanently separate predictor across the
season.

SOLUTION

The prediction league table (Chapter 8) ranks a small, fixed set of
real, permanent predictors across the whole season: the user, the BBC
expert, "guest," and the BBC's AI prediction. A guest, by contrast, is
genuinely transient — some weeks have no guest at all, some weeks have
one, and some weeks have several different people, none of whom
appear in every other gameweek.

If each individual guest were tracked as their own permanent row in
the prediction league table, the table would end up full of
predictors who only ever made one or two predictions all season,
which:

  1. Makes the season-long ranking meaningless for them — a single
     lucky or unlucky guess produces a wildly unrepresentative
     "average," since there's no larger sample of predictions to
     smooth it out.
  2. Clutters the table with names that never appear again, instead
     of keeping it focused on the predictors who are actually being
     compared across the full 38-gameweek season.

Averaging every guest's own score into one shared "guest" slot for
that week solves both problems: it keeps the four real predictor
categories fixed and comparable across the whole season, while still
giving genuine weight and visibility to however many guests happened
to take part in any individual week.

WHY THIS WORKS AS AN ANSWER
----------------------------
It identifies the real underlying problem (an unstable, ever-changing
set of one-off predictors polluting a season-long ranking table) and
explains concretely why averaging into one shared slot is the fix,
rather than just restating that averaging happens.