Premier League Predictor: FastAPI & PostgreSQL — Chapter 7, Exercise 3
====================================================

TASK
Enter real results for at least three fixtures in a real season (some
involving the same team as both home and away across different
gameweeks), call GET /api/seasons/{id}/table, and confirm by hand that
at least one team's played/won/drawn/lost/points figures correctly
combine both its home and away results.

SOLUTION

Using a real season with Team A appearing in two different fixtures —
once at home, once away — plus a third unrelated fixture:

1. Gameweek 1: Team A (home) 2-1 Team B (away) — PATCH the result via
   /api/fixtures/{id}/result.
2. Gameweek 2: Team C (home) 0-0 Team A (away) — PATCH this result too.
3. Gameweek 2: Team D (home) 1-3 Team E (away) — a third, unrelated
   result, entered just to confirm the table handles more than one
   team correctly at once.

By hand, Team A's real combined record across its two fixtures:
- Fixture 1 (home): won 2-1 → 1 win, goals_for 2, goals_against 1,
  3 points.
- Fixture 2 (away): drew 0-0 → 1 draw, goals_for 0, goals_against 0,
  1 point.
- Combined: played 2, won 1, drawn 1, lost 0, goals_for 2,
  goals_against 1, goal_difference +1, points 4.

Calling GET /api/seasons/{id}/table and finding Team A's row should
show exactly played: 2, won: 1, drawn: 1, lost: 0, goals_for: 2,
goals_against: 1, goal_difference: 1, points: 4 — matching the by-hand
calculation exactly, with the away 0-0 draw correctly folded in
alongside the home win rather than only one of the two fixtures being
counted.

Team D and Team E's rows should independently show Team D with 1
played, 0 won, 0 drawn, 1 lost, goals_for 1, goals_against 3, goal
difference -2, 0 points; and Team E with 1 played, 1 won, 0 drawn, 0
lost, goals_for 3, goals_against 1, goal difference +2, 3 points —
confirming the query correctly separates each team's own record rather
than mixing fixtures together.

WHY THIS WORKS AS AN ANSWER
----------------------------
It deliberately constructs a scenario where the same team plays once at
home and once away, computes that team's real combined record by hand
first, then verifies the query's actual response matches that
hand-calculated result exactly — directly testing the UNION ALL
normalization from Exercise 1 rather than just trusting it works.