Spaced repetition is a review schedule that spreads study sessions across widening time gaps instead of one long block. In a 2006 meta-analysis of 317 experiments, distributed practice beat cramming for long-term retention by 10 to 30 percent across every age group and subject tested.
The method is old. Hermann Ebbinghaus mapped the first forgetting curve in 1885, testing his own recall of nonsense syllables over hours and days.
What is new is the precision. Modern algorithms now predict the day a specific fact will slip, then set the review for the moment just before it does.
What Is Spaced Repetition?
Spaced repetition is a study technique where you review information at increasing intervals, revisiting each fact just before you are likely to forget it. Get a recall right, and the next review moves further out. Get it wrong, and the fact comes back sooner.
The engine behind it is active recall. Every review forces you to retrieve the answer from memory rather than reread it, and retrieval is what strengthens the trace.
A concrete example: you learn 20 French words on Monday, review them Wednesday, then the following Monday, then two weeks later. That widening ladder is the whole idea.
Why Does Spacing Beat Cramming?
Cramming loads everything into one session. Spacing splits the same material across sessions with gaps between them. On any test more than a day out, the split version wins.
The evidence is among the most replicated in cognitive psychology. Ebbinghaus documented the forgetting curve in 1885, and Herbert Spitzer confirmed the classroom version in 1939 with 3,605 sixth-graders across Iowa.
The modern synthesis is Cepeda and colleagues' 2006 meta-analysis in Psychological Bulletin, which pooled 317 experiments and found distributed practice beat massed practice by 10 to 30 percent for long-term retention.
The reason is not effort. It is timing. Each time you let a memory fade slightly and then retrieve it, the retrieval does more work than an easy one, and the memory returns more stable. This is why cramming the night before an exam feels productive but rarely survives the week.
How Does Spaced Repetition Work?
Two rules run the whole system. Widen the gap after a success. Shorten it after a failure. Everything else is bookkeeping.
The gap matters because memory decays on a curve, not a line. You forget fastest right after learning, then more slowly. Ebbinghaus's data put roughly 67 percent of new material gone within a day without review.
Reviewing just before that drop resets the curve. The memory then fades more slowly the next time, which is why intervals keep stretching from days to weeks to months.
A 2015 replication by Murre and Dros reproduced Ebbinghaus's curve almost exactly, with one twist: a small recovery at the 24-hour mark they linked to a night of sleep. Sleep consolidates memory, so an overnight gap does real work.
What Are The Main Methods?
Three systems dominate, and they differ mainly in how they pick the next interval. The oldest needs no computer. The newest fits a statistical model to your personal history.
The Leitner system, introduced by Sebastian Leitner in 1972, uses physical boxes. A correct card moves up to a box with a longer interval; a wrong card drops to box one. Simple, tactile, fine for small decks.
SM-2 came next. Piotr Woźniak published it in 1987 for SuperMemo, giving each card an "ease factor" that starts at 2.5 and stretches the interval on every success. It powered Anki from 2006 until late 2023.
FSRS is the current state of the art. Built by Jarrett Ye and the open-source community since 2022, it models three variables per card, difficulty, stability, and retrievability, and predicts your recall probability directly. Anki made it the default in November 2023.
On benchmarks across more than 500 million real Anki reviews, FSRS needs 20 to 30 percent fewer reviews than SM-2 to hit the same retention. The gap is scheduling accuracy, not effort.
Most study software today runs one of these three schedulers. Anki and RemNote offer FSRS; SuperMemo runs its own proprietary line; Quizlet-style apps use a Leitner-box approximation. AskSia's flashcard maker defaults to FSRS with SM-2 as a fallback.
What Do 2-3-5-7 And 2-7-30 Mean?
Searches for spaced repetition surface two "magic number" schedules more than any others: 2-3-5-7 and 2-7-30. Both are manual routines, not algorithms, and neither comes from a specific study.
They are shorthand. Each is a fixed ladder of review gaps you follow by calendar, with no software deciding anything. That makes them easy to start and easy to mistake for science.
Here is what most guides get wrong. These fixed schedules treat every fact the same. A word you know cold and a word you keep missing both get identical spacing, so you overstudy the easy one and underschedule the hard one.
The real principle they approximate is proportional, not fixed. Cepeda's 2008 study showed the ideal gap scales to how long you need the memory, which no single 2-3-5-7 ladder can match across different material.
How Do You Build A Schedule?
Start with your target. The gap between reviews should track how long you need to remember something, not a fixed rule.
Cepeda and colleagues' 2008 study of 1,354 learners gave the working number: space reviews at roughly 10 to 20 percent of your retention window.
For a final six weeks out, that means first reviews a few days apart, widening as the exam nears. For vocabulary you want to keep for a year, reviews start weeks apart. To fold this into a broader routine, see our guide on how to study effectively.
Doing this by hand works up to a point. Past a few dozen items, tracking which card is due on which day becomes its own job, which is where software earns its place.
Two habits decide whether it works. Review daily so the due pile never snowballs, and make every review a real retrieval attempt, not a reread. Pairing spaced cards with an active quiz format, like AskSia's quiz generator, keeps the retrieval honest.
Will Spaced Repetition Always Work?
Spaced repetition builds recall, not understanding. It is strong for facts, terms, formulas, and vocabulary. It does little for a concept you never grasped in the first place.
A flashcard can drill the definition of opportunity cost. It cannot teach you to reason with it. For that you need worked problems first, then cards to lock in the pieces.
The second failure is upkeep. The system only works if you show up daily. Skip a week and the due pile balloons, which is the most common reason students quit.
The third is the deck itself. Hand-building hundreds of good cards can take longer than the studying, which is why generating cards from your own material removes the step that stops most people before they start.
Frequently Asked Questions
What is an example of spaced repetition?
A student learning 30 biology terms reviews them one day after class, then three days later, then a week later, then two weeks on. Each time recall succeeds, the next gap widens. That expanding schedule is spaced repetition in its simplest form. A flashcard app does the same automatically: Anki, which has used spaced-repetition scheduling since its 2006 launch, shows a card, you rate how well you recalled it, and the algorithm sets the next date. Language learners are the classic case. Reviewing 20 vocabulary words across five short sessions over two weeks produces far stronger retention than one 100-minute block, because each spaced retrieval forces your brain to reconstruct the memory rather than recognize it. To try it on vocabulary, run a word list through WordJump and let the schedule handle the timing.
What is the 2-3-5-7 method?
The 2-3-5-7 method is a fixed manual schedule: review new material 2 days after learning it, again after 3 more days, then 5, then 7, each number counting the gap since the last review. It is a simplified stand-in for true spaced repetition, popular because it needs no software, only a calendar. The intervals are not drawn from any single study. They approximate the core finding from Cepeda and colleagues' 2006 meta-analysis: gaps should widen after each successful recall. The method's weakness is that it treats every fact identically. A term you find easy and a term you keep failing both get the same 2-3-5-7 spacing, which wastes reviews on the easy one and under-schedules the hard one. For a small set of notes it works fine. For hundreds of cards, use an adaptive app instead.
What is the 2-7-30 rule?
The 2-7-30 rule schedules three reviews after the first study session: one at 2 days, one at 7 days, one at 30 days. Like 2-3-5-7, it is a manual heuristic rather than a lab result, and the numbers are chosen for memorability as much as accuracy. It maps loosely onto the proportional spacing principle from Cepeda et al.'s 2008 study of 1,354 learners, which found the ideal gap runs roughly 10 to 20 percent of your target retention window. For a fact you need to hold for about a year, three reviews across the first month is a reasonable opening sequence, though a real algorithm would keep extending the gaps well past day 30. Treat 2-7-30 as a starting scaffold for material you study by hand, then graduate to an app once your review pile outgrows a calendar.
What are the benefits of spaced repetition?
The headline benefit is durability. Cepeda et al.'s 2006 meta-analysis of 317 experiments found distributed practice beat massed practice for long-term retention by 10 to 30 percent, and a 2008 follow-up put the per-hour gain even higher on delayed tests. The second benefit is efficiency. Because you review a fact only when you are about to forget it, you spend fewer total minutes than re-reading everything repeatedly. On mature decks, FSRS cuts review volume by 20 to 30 percent against the older SM-2 for the same retention. The third is exam transfer: spacing pairs naturally with active recall, one of the most reliable boosts to test performance in the research. The catch is that benefits show up on delayed tests, not next-morning quizzes, so start at least two weeks out. Build it into your exam study routine rather than as a separate chore.
What software uses spaced repetition?
The best-known tool is Anki, free and open-source, which ran SM-2 from 2006 and added FSRS as its default scheduler in version 23.10 in November 2023. SuperMemo, the original 1987 application, still runs its own proprietary algorithms. RemNote and dozens of newer apps also offer FSRS, while Quizlet-style apps use a simpler Leitner-box approach rather than true adaptive scheduling. AskSia's flashcard maker runs FSRS by default and generates the cards from your uploads, so you skip the hours of manual typing that stop most students. Which one wins depends less on the algorithm than on whether you open it daily, since the largest gap in the research is between using spaced repetition at all and not. For a side-by-side on the main options, see our roundup of the best study apps.
Is spaced repetition better than cramming?
For anything you need past tomorrow, yes, and the margin is large. Cramming reliably produces short-term recognition and long-term loss. Ebbinghaus's 1885 data showed roughly 67 percent of freshly learned material gone within 24 hours without review, and Murre and Dros reproduced that curve almost exactly in 2015. Spaced repetition beats it by 10 to 30 percent on delayed tests. The one case where cramming holds up is a low-stakes quiz the very next morning, where there is no time for spacing to pay off and retention past the test does not matter. For a cumulative final, a licensing exam, or any material you will use again, spacing wins decisively. If you have only one night left, our guide on what to do the night before an exam covers the damage-control version.
