Module 3 — Spaced Repetition

Builds on Modules 1–2 · hands-on · about 25 minutes.

You now know that retrieval practice beats passive review. But when should you retrieve? It turns out the timing matters enormously — and the best timing is counterintuitive: you should wait until you are just about to forget before you practice again. Space your reviews out, and each one does far more work than if you had massed them together. This is the spacing effect, one of the most durable findings in all of memory science.

Massed practice vs. spaced practice

Imagine you have one hour to learn a set of vocabulary words before a test in four weeks. You could spend the whole hour today — what we call massed practice or cramming. Or you could spread that hour across four 15-minute sessions on, say, days 1, 7, 21, and 42. For the same total investment of time, the spaced schedule produces dramatically better retention weeks later.

This is the core of the spacing effect (first documented by Ebbinghaus in the 1880s, replicated hundreds of times since): distributing practice over time beats concentrating it into a single block, even when the total practice time is identical. Cramming works fine for a test tomorrow — but the material evaporates almost as quickly as it was loaded. Spaced practice builds memories that last.

In terms of the forgetting-curve model from Module 1, here is what happens: each review session resets retention toward 1 and, because successful retrieval strengthens the memory, multiplies the strength \( S \) — so the curve following each review is flatter than the one before it. When you space reviews out, you catch each one just before forgetting, and each strengthening compounds on the last. Massed practice piles reviews together so the later ones find a memory that is still near full strength — they do far less work.

Why cramming feels like it works The night before an exam, cramming does raise your score — because the test happens while the memory is still warm. The problem is what happens next. Studies tracking retention weeks after an exam find that crammers' scores collapse far faster than spaced learners' scores. Cramming optimizes for the test date, not for durable knowledge. If you are learning something you actually want to keep — a language, a professional skill, a medical concept — cramming is close to worthless.

Spaced repetition systems

The spacing effect is a general principle; a spaced repetition system (SRS) is a precise implementation of it. The key idea is that each item in your study deck has its own scheduled review date, and that date is computed from your past performance on that item. After a successful recall, the item's next review is pushed further into the future — because you have demonstrated you can remember it, so it does not need attention soon. After a lapse, the interval resets short — because clearly you need more practice.

The most influential algorithm is SM-2, developed by Piotr Wozniak and the basis of the popular flashcard app Anki. In simplified form:

Each card has a current interval (number of days until the next review) and an ease factor (a multiplier, starting around 2.5).
On a successful recall: new interval = round(current interval × ease). So a card you knew well might go from 5 days → 11 days → 24 days → 53 days. The exponential growth means you eventually need to review a well-learned card only once a year.
On a lapse: the interval resets to 1 day (or a short "relearn" period), and the ease factor decreases slightly — the system recognizes this item is harder for you than it thought.

The result is a schedule that keeps each item at the edge of forgetting: reviewed often when it is new and fragile, reviewed rarely when it is consolidated and strong. This is sometimes called studying at the forgetting threshold — the single most efficient point to review.

Real world: Anki and the medical profession Medical students face the most extreme version of the retention problem: thousands of facts, drugs, dosages, and mechanisms that must be recalled reliably under pressure, years after learning. Anki — a free, open-source SRS — has become close to standard practice in many medical schools. Students who use it consistently report needing far less last-minute cramming before clinical rotations, because the knowledge was maintained continuously. Language learners use similar tools: Duolingo's algorithm schedules vocabulary review using spacing principles, and dedicated apps like Clozemaster are built entirely around SM-2-style scheduling.

Massed vs. spaced — see the difference

The activity below models two learners who each review a memory three times, but on different schedules. Retention follows \( R(t) = e^{-t/S} \), and each review resets retention to 1 and multiplies strength by a boost factor. The spaced learner reviews on days 0, 7, and 21; the massed learner reviews on days 0, 1, and 2. Watch day-30 retention for each.

This activity needs JavaScript. The idea: three reviews massed together barely improve long-term retention because later reviews find the memory still strong; spaced reviews compound more because each one arrives when forgetting has reduced retention.

Build your own spaced-repetition schedule

Step through a single flashcard's journey. Click "Recalled" when you would have remembered it, or "Forgot" when you would not have. Watch how the interval grows on success and snaps back on a lapse — that is the SM-2 logic in action.

This activity needs JavaScript. It simulates an SRS scheduler: successful recalls grow the interval, lapses reset it.

Why this matters next Spaced repetition solves the when problem for a body of material you already have. But learning also requires a bar: how much is "good enough" before you move on to the next concept? And what should happen for learners who need more time than others? That is mastery learning (Module 4) — the principle that the learning standard should be held fixed while the time each learner needs is allowed to vary.

One-sentence summary: for the same total study time, spreading practice across multiple sessions produces far more durable memory than massing it together, because each spaced review arrives as the trace is fading and does the maximum consolidation work — and a spaced-repetition system automates this scheduling per-item, growing the interval after success and resetting it after a lapse.

Next: Mastery Learning & the Zone →