Spaced Repetition, Explained Properly
2026-08-02 · 8 min read
Why reviewing at increasing intervals beats cramming, what the forgetting curve actually shows, and how to build a schedule without an app.
Forgetting is normal and predictable In the 1880s Hermann Ebbinghaus memorised long lists of nonsense syllables and measured how quickly he lost them. The result — the forgetting curve — is steep at first and then flattens. Most of what you lose, you lose in the first day or two.
The important follow-up finding is what happens when you review. Each successful review flattens the curve. After the first review you forget more slowly; after the third, slower still. The interval you can leave between reviews grows roughly geometrically. That is the whole idea behind spaced repetition.
Why cramming feels better and works worse Cramming produces high performance immediately and poor performance later. Massed practice keeps material in an easily accessible state, so a test the next morning may go fine. A cumulative final three weeks later will not.
Spacing feels worse while you do it. You return to material you have partly forgotten, and retrieving it is effortful. That effort is the mechanism — the difficulty of the retrieval is what strengthens the memory. Psychologists call this desirable difficulty, and it is one of the most robust findings in the field.
A schedule you can run on paper You do not need software. A basic schedule that works for most school material:
- Review 1: the same day you learn it, ten minutes, from memory.
- Review 2: the next day.
- Review 3: three days later.
- Review 4: one week later.
- Review 5: three weeks later.
Each review should be a retrieval attempt, not a reread. Close the book, write what you remember, then check. If a review goes badly, drop back one interval instead of continuing to expand.
Keep a simple index card box with five dividers labelled by interval, or a page in a planner with dates. The mechanism matters more than the tooling.
What belongs in a spaced system Facts, vocabulary, formulas, dates, definitions, sequences, and small procedures. Anything with a clear prompt and a clear answer.
What does not belong: essay writing, long proofs, lab technique, and anything whose "answer" is a judgement rather than a fact. Those need practice and feedback, not flashcards. A common failure is trying to reduce an entire history course to cards; you end up memorising dates while remaining unable to construct an argument, which is what the exam actually asks for.
Writing cards that work A good card has one idea, an unambiguous prompt, and an answer you could say out loud in under fifteen seconds. "Explain World War I" is not a card. "What alliance obligation pulled Russia into the war after Austria-Hungary's ultimatum to Serbia?" is.
Avoid cards whose answer is a list of eight things. Split them. Lists are recalled poorly and give you no partial credit against yourself.
Include the "why" where you can. A card that reads "mitosis produces two identical diploid cells — because the goal is growth and repair, not variation" is retained better than the bare fact, because the reason gives the fact somewhere to attach.