Premier League Predictor: FastAPI & Redis — Chapter 6, Exercise 2
=================================================================

TASK
Write a second version of score_prediction that checks the outcome before
the exact score. Run both versions on five cases (including two draws)
and report which cases differ and why.

SOLUTION

  def swapped(pred, actual):
      if sign(*pred) == sign(*actual):   # outcome first
          return 10
      return 40 if pred == actual else 0

Cases (predicted, actual):
  (0-0, 0-0)   (1-1, 2-2)   (2-1, 1-2)   (2-1, 2-1)   (3-0, 1-0)

Real results:

  correct order:  [40, 10, 0, 40, 10]
  swapped order:  [10, 10, 0, 10, 10]

Cases 1 and 4 differ: 0-0 vs 0-0 and 2-1 vs 2-1 are exact matches, and
the swapped version scores them 10 instead of 40. An exact score always
also has the right outcome, so the outcome check fires first and returns
before the exact-match check is ever reached. The other three cases agree
because none of them is an exact match, so the order can't matter.

Draws are not special: 1-1 vs 2-2 scores 10 in both versions because the
outcome (a draw) matches while the scoreline does not.

WHY THIS WORKS AS AN ANSWER
----------------------------
It reproduces the bug with real output, identifies precisely which cases
are affected (exact matches only), and states the general rule: when one
condition implies another, test the more specific one first.