Safe demo rules for your toy model

Skip production-ops buzzwords. Use school-clear gates:

1. Don’t share real classmate messages — use made-up practice texts or sports scores you invented.
2. Don’t claim your toy model works on all phones — you only proved five fixture rounds.
3. Write what you proved vs what you didn’t — perfect accuracy here is not a worldwide SLA.
4. Don’t hide false alarms — always show TP/FP/TN/FN with accuracy.

If any gate fails, the demo stays offline until fixed.

Your evidence for “the math works on this practice set” stays:

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

That line may appear on the demo board under a sticky note: Proved on n=5 practice rounds only.

Practice saying: “This toy turns scores into yes/no with a cutoff. On five labeled rounds it matched every guess. It does not prove it catches every spam text at school or every free throw in the gym.”

If someone asks “So it’s ready for the school app?” your gate answer is: “Not with n=5 — that would be overclaiming.”

Name who can stop the demo (you + a teacher/club lead). Name what you delete after demo day (any real messages if someone slipped). Keep rollback simple: close the laptop, put the poster away, revert to the fixture-only notebook.