Spaced Repetition, Explained: The Method Behind Long-Term Memory

Spaced repetition is a review method that spreads study sessions across increasing intervals of time instead of repeating material in one sitting. It works because memory decays predictably, and reviewing a fact just before you would forget it strengthens the memory more than reviewing it while it is still fresh.

Short answer

• Spaced repetition schedules reviews at increasing intervals instead of all at once.
• It works because of the forgetting curve: memory fades fast at first, then slower with each successful review.
• An algorithm decides the interval for each card based on how hard it was to recall and how long ago you last saw it.
• Cramming feels productive but produces weaker long-term memory than the same time spent spaced out.
• Modern schedulers like FSRS replace fixed intervals with a memory model built from your own review history.

What is spaced repetition?

Spaced repetition is a learning technique where you review the same piece of information multiple times, with the gap between reviews growing each time you recall it correctly. Instead of studying a fact once and moving on, or re-reading it ten times in one session, you see it again right around the point where you would otherwise forget it.

The technique applies to anything you need to recall on demand: vocabulary, anatomy terms, historical dates, chemical formulas, or the wiring diagram you memorized for a certification exam. It does not help with skills that depend on practice and feedback rather than recall, like playing an instrument or writing an essay.

Three ideas make the method work together:

1. Forgetting is predictable. Without review, recall probability drops along a curve, not randomly.
2. A successful recall resets the clock and lengthens it. Each time you retrieve a fact correctly, the next safe gap gets longer.
3. A failed recall shortens the interval. A card you get wrong needs to come back sooner, not later.

Flashcards are the most common delivery format because each card isolates one fact and gives a clear pass or fail signal the scheduler can use. That signal is what separates spaced repetition from ordinary flashcard use: the schedule adapts to your answers instead of following a fixed order.

The forgetting curve: why review timing matters

Memory researcher Hermann Ebbinghaus ran the first systematic study of forgetting in the 1880s, testing his own recall of nonsense syllables over days and weeks. His results, still cited in memory research today, showed that most forgetting happens fast: a large share of newly learned material is lost within the first day, and the rate of loss slows down after that. This decay pattern is what people mean by "the forgetting curve."

The practical consequence is that a single review, done at the right moment, recovers most of what a second read-through would have. Reviewing too early wastes time on a memory that has not decayed yet. Reviewing too late means relearning something you already forgot, which costs more effort than a well-timed nudge would have.

A 2006 meta-analysis by Cepeda and colleagues, published in Psychological Bulletin, reviewed decades of spacing-effect research and confirmed the pattern across many kinds of material: distributing study sessions over time produced better long-term retention than massing them together, and the ideal gap between sessions grows as the total time you plan to remember something grows.

This is the mechanism spaced repetition automates. Instead of you guessing when a memory is about to fade, a scheduler tracks it per card and tells you when to look again.

How a spaced repetition algorithm decides your next review

An algorithm's job is to pick the review interval that gives you the highest chance of a successful recall without wasting a review on a card you would have remembered anyway. Every scheduler answers the same question — "when should this specific card come back?" — but the methods have gotten more precise over four decades.

The oldest widely used method, the Leitner system, sorts physical cards into boxes: a correct answer moves a card to a box reviewed less often, a wrong answer sends it back to the first box. It requires no computer and no statistics, and it is still a reasonable way to organize paper flashcards.

Software schedulers replaced fixed boxes with a formula. The SM-2 algorithm, published by Piotr Woźniak in 1987 as part of the SuperMemo method, is the ancestor most flashcard software has used in some form since. It keeps an "ease factor" per card that grows when you recall correctly and shrinks when you don't, and multiplies the previous interval by that factor to get the next one.

SM-2 was a large improvement over fixed steps, but it has a known weakness: it reacts to your last answer, not to the shape of your memory. Two cards with the same ease factor can behave very differently depending on how naturally memorable they are, and SM-2 has no separate way to represent that.

FSRS (Free Spaced Repetition Scheduler) is an open-source algorithm built to fix that gap. Instead of one ease number, it models two things separately for every card: stability, how long the memory is expected to last before recall probability drops to your target level, and difficulty, how hard that specific card tends to be for you regardless of timing. The model is fitted to your own review history, so two people studying the same deck end up with different schedules.

For the mechanics without the underlying math, see how FSRS actually schedules a card.

Spaced repetition vs. cramming: what the research shows

Cramming produces recognition, not durable recall. You can push a fact into short-term memory through repetition in one sitting and pass a quiz an hour later, but that same fact is often gone within days because it never went through the retrieval-and-regap cycle that builds long-term storage.

The Cepeda meta-analysis cited above is one of the clearer demonstrations of this gap: across the studies it reviewed, spaced study sessions consistently outperformed massed study on delayed tests, and the advantage grew larger the longer the delay between learning and the final test. Cramming can match spaced study on a test given the next day. It falls behind on a test given a month later, which is closer to how most real exams and real jobs actually test you.

There is a second, less discussed cost to cramming: it hides what you don't know. Re-reading notes or highlighting a textbook creates a feeling of familiarity that people mistake for mastery. Active recall — trying to produce the answer before you see it — exposes the gap. Spaced repetition forces active recall by design, because every review starts with the question, not the answer.

For a closer look at why recall beats re-reading specifically, see active recall vs. re-reading.

What makes a review interval "correct"

There is no universal correct interval — only an interval that matches the retention you're aiming for. A scheduler like FSRS lets you set a desired retention, the probability you want to still recall a card at the moment it comes due. A higher target (say, 95%) produces shorter, more frequent reviews. A lower target (say, 80%) produces longer, less frequent reviews and more forgetting along the way.

Three factors combine to set each card's actual interval:

• Stability: how long this specific card's memory is expected to hold at your chosen retention target.
• Difficulty: how much harder or easier this card is for you than an average card, independent of timing.
• Elapsed time since the last review: the algorithm compares how long it has actually been against how long it predicted, and adjusts.

This is why two cards reviewed on the same day can get wildly different next intervals. A card you have answered correctly ten times in a row with high confidence might not come back for months. A card you keep missing might return the next day even if it "should" be due later on a fixed schedule.

For the specific trade-off between studying less and forgetting more, see choosing a desired retention target.

Common mistakes that break spaced repetition

The method fails less often because of the algorithm and more often because of how it gets used. Four mistakes show up repeatedly.

Adding too many new cards per day

Every new card becomes a future review. Adding fifty new cards a day compounds into hundreds of daily reviews within a few weeks, and the backlog itself becomes the reason people quit. The fix is a daily new-card limit sized to what you can actually review going forward, not to how motivated you feel today. See how many new cards to add per day.

Grading recall by feeling instead of by producing the answer

Spaced repetition depends on an honest pass or fail signal. Looking at a card, thinking "yeah, I know this," and flipping straight to "correct" without producing the answer first breaks the model: the algorithm schedules the next review as if you actually recalled it, when you only recognized it.

Editing cards without expecting a reset

Rewriting a card's question changes what you're actually being asked to recall. A scheduler that has built up months of stability for the old question can't fully carry that trust to a materially different one. Minor wording fixes are fine; changing what the card asks is closer to a new card.

Cramming before the exam instead of trusting the schedule

The instinct to binge-review everything the night before an exam undoes the spacing that built the memory in the first place, and it adds fatigue right when you need to perform. If a backlog has piled up before a deadline, working through it in a structured way beats a random cramming sprint. See how to recover from a flashcard backlog for a recovery plan that doesn't require starting over.

How to start a spaced repetition practice this week

You don't need a perfect deck or a fully tuned algorithm to start. You need a small set of cards, a daily habit, and honest grading.

1. Pick one real subject, not a demo deck. A course you're taking, a language you're learning, or material for a certification exam works better than generic trivia, because you already have a reason to remember it.
2. Write your own cards for the first week. Cards you phrase yourself are easier to recall later than cards copied from someone else's deck, because the act of writing them is itself a first exposure.
3. Set a small daily new-card limit — 10 to 15 is a reasonable start — and let the scheduler build your queue from there instead of importing thousands of cards at once.
4. Review at the same time each day. Consistency matters more than session length for a method built around timing.
5. Grade honestly. Try to produce the answer before flipping the card. A generous grade today creates a harder review later.
6. Check the interval it gives you for a card you found easy versus one you found hard. Watching the numbers change based on your own answers is the fastest way to trust the method.

A three-card test is enough to see the mechanism work: create three cards, get one right, one wrong, and one uncertain, and compare the next intervals the scheduler assigns to each. Open StudyDaily in your browser to try this without installing anything — reviewing doesn't require an account.

How StudyDaily implements spaced repetition

StudyDaily schedules every card with FSRS-6, the current version of the open-source algorithm described above. Each rating button — Again, Hard, Good, Easy — shows you the interval it will produce before you tap it, so the schedule stays visible instead of hidden behind a black box.

The scheduler tracks stability and difficulty per card, fitted to your own review history rather than a generic curve. Reviews, card state, and statistics all run on the device: nothing about your daily study depends on a network connection or an account. An account exists only for two things — AI card generation and cloud backup — so the core spaced repetition workflow works the same with or without one.

I built the scheduler around a simple test: show the interval, then trust the person reviewing to judge whether it feels right for that card. That is why the rating buttons expose the number instead of hiding a decision the reader can't see or question.

What should you do next?

Start with one deck you actually need, a small daily new-card limit, and honest grading. Spaced repetition rewards patience with a schedule more than it rewards intensity in a single session.

• Write or import a small, real deck instead of a demo.
• Cap new cards per day at a number you can sustain.
• Review at the same time daily and grade before you see the interval.
• Watch how the schedule changes for cards you find easy versus hard.
• Trust the spacing when an exam approaches instead of cramming against it.

Questions people ask

Does spaced repetition actually work?

Yes. Spacing-effect research going back to Ebbinghaus and confirmed by later meta-analyses, including Cepeda et al. (2006) in Psychological Bulletin, consistently shows spaced study producing better long-term retention than massed study or cramming, with the advantage growing on delayed tests.

How is FSRS different from older algorithms like SM-2?

SM-2 tracks a single ease factor per card and multiplies the previous interval by it. FSRS models two separate properties per card — stability and difficulty — fitted to your own review history, which lets it predict recall probability more precisely than a one-number formula can.

How often should I review a flashcard?

It depends on the card, not a fixed rule. A well-timed scheduler reviews an easy, well-known card only occasionally and a hard or recently missed card much sooner. Setting a desired retention target lets you choose the general trade-off between review frequency and forgetting.

Can spaced repetition replace studying entirely?

No. It is a retention method for material you already need to recall on demand, not a substitute for understanding a concept the first time. Read or work through new material first, then use spaced repetition to keep it retrievable over time.

Sources used for this guide

Ebbinghaus's original forgetting-curve study dates to the 1880s and is summarized in most cognitive psychology textbooks; the spacing-effect meta-analysis by Cepeda et al. (2006) in Psychological Bulletin provides the modern, peer-reviewed confirmation cited above. Algorithm history draws on the SuperMemo SM-2 method description and the FSRS project documentation, which is open source and describes the stability/difficulty model in detail. StudyDaily's scheduling claims come from the product's own FSRS-6 implementation, checked against the sources above for this guide's 8 June 2026 publication date.

SuperMemo is a registered trademark of its respective owner. StudyDaily is not affiliated with, endorsed by, or sponsored by SuperMemo. The product name is used here only to describe the history of spaced repetition algorithms.