Finals week starts with a decision most students get wrong: where to begin. The default is rereading from lecture 1 — start at the beginning, work forward, cover everything. It feels thorough and it guarantees you'll run out of time in the wrong place: deep into the early lectures you mostly knew, out of hours when you reach the weak ones at the end.

The fix is a diagnostic before the review. Sample every lecture with two closed-book prompts, score them, and let the misses rank the repair list. Only then do you reopen notes — and only for the lectures the diagnostic flagged. Triage first, repair second.

The diagnostic's job isn't studying; it's finding where studying is needed. A scored miss is a gap; "I kind of remember that" is the same gap wearing fluency's clothes. The mixed closed-book set is what separates them.

Why rereading the whole folder fails

Rereading from lecture 1 is the finals-week failure mode: it's sequential, it's passive, and it hides the gaps it should find. You start fresh on material you already half-know, spend hours feeling productive, and reach the lectures you actually need when the week is almost gone. Coverage-by-rereading is how students study the wrong 70% — the comfortable 70% that didn't need the time.

Retrieval makes gaps visible; restudy does not. When you reread, the material is right there and "I remember this" is indistinguishable from "I can produce this." A closed-book prompt forces the distinction — you either produce the answer or you don't, and the don't is the diagnosis. The mixed set also practices the discrimination finals require: questions drawn from across the term, not sorted by chapter, which is the exam's own structure.

The diagnose-then-repair workflow

Sample the whole term. Two prompts per lecture, covering every lecture the course had — not a deep dive, a survey. The diagnostic's job is mapping the territory: which lectures hold, which have gaps. Two items per lecture is enough to rank without becoming a study session.

Score each attempt. Not "I remember that" — produce the answer, then score it against the notes. 2/2 means solid; anything less is a flag. The score is what separates "feels familiar" from "can retrieve," which is the distinction that matters.

Rank the misses into a repair list. Worst first, not chronologically. Lecture 3 with two misses goes above lecture 1 with none — even though lecture 1 comes first in the folder. The rank is the week's plan; the folder order isn't.

Reopen notes only for flagged lectures. The diagnostic tells you where to repair; the repair sessions are for the misses, not the whole folder. A lecture that scored 2/2 doesn't get re-studied — it gets a spaced re-test, which is a different, shorter task.

Retest with new stems. Same lecture, different question — the 1/3/7-day spacing holds, and the new stem prevents the diagnostic items themselves from becoming memorized answers instead of tested knowledge.

A 40-minute session on the clock

The weak answer is the diagnostic's target: "I remember they disagreed" is the feeling of familiarity without the retrieval — the student knows there's a contrast but can't produce it. The strong answer produces both theories' positions and, crucially, names what's missing: the concrete policy examples. That "what's missing" is the repair list — the specific content that needs work, not the whole lecture.

Why the diagnostic comes first

The diagnostic and the reread answer different questions. The reread asks "what does this material say?" — a question the exam doesn't ask. The diagnostic asks "what can I produce?" — which is exactly the exam's question. Starting with the reread answers the wrong question across the whole folder; starting with the diagnostic answers the right one and tells you where the wrong one needs filling.

This is why the diagnostic caps at 45 minutes. Its job is ranking, not repairing — the moment it becomes a study session, it's rereading with extra steps. Two prompts per lecture, score them, rank the misses, stop. The repair comes after, and only where the diagnostic pointed.