Set the rules of the guess lists

Think of the code as a board game with three piles:

1. Truth pile — only 0 or 1 for each round (did the shot go in? was it spam?).
2. Score pile — numbers between 0 and 1 (how strong the clue looked).
3. Threshold — one cutoff, also between 0 and 1.

Same length for truth and scores. No secret third list. No pasting real classmate messages into the demo — use made-up practice rounds.

If a score is out of range or the piles do not line up, that is a rules bug, not proof that “the AI is dumb.” Fix the rules before you trust the scoreboard.

We still run the happy fixture so you can see a clean contract in action:

truth = [1, 0, 1, 1, 0]
scores = [0.8, 0.3, 0.6, 0.9, 0.2]
threshold = 0.5
preds = [int(score >= threshold) for score in scores]

pairs = list(zip(truth, preds))
tp = pairs.count((1, 1)); fp = pairs.count((0, 1))
tn = pairs.count((0, 0)); fn = pairs.count((1, 0))
accuracy = (tp + tn) / len(pairs)
print(preds, f"accuracy={accuracy:.2f}", tp, fp, tn, fn)

Expected evidence:

[1, 0, 1, 1, 0] accuracy=1.00 3 0 2 0

• Every score sits in [0, 1], and so does threshold.
• truth and scores both have length 5, so zip pairs them honestly.
• Guesses are only 0 or 1, never a mystery third label.
• The printout names predictions and the four buckets, so nobody can hide behind a lonely accuracy number.

Write a one-line rule for each field: required shape, allowed range, and what is forbidden (real phone numbers, real chat logs, scores above 1).

First make sure the piles are legal. Only then count TP/FP/TN/FN. Mixing those jobs makes debugging feel like random blame: Was the threshold wrong, or did someone paste six scores and five truths?