The questionable claim
A learner tries twenty model settings and chooses the one with the best final test score.
Investigate
Identify the unit of analysis, denominator, population or information boundary. What extra evidence would be needed? Write a corrected statement first.
Diagnosis and repair
The test set has influenced selection and is no longer an untouched estimate. Use training/validation or cross-validation for selection, then evaluate once on a held-out set. Disclose previous reuse.
Prevention
Add a test or review question that catches the error before publication. Save the original and corrected explanation so your reasoning is auditable.
Try the activity before reading its reference. Keep your sources, commands and assumptions with your work. Companion activities are unsaved self-practice and do not award a certificate or change lesson progress.