Spaced Repetition Explained: How It Actually Works (and Why)

Author: Anatole Gaigneux · Published: 2026-04-30 · Updated: 2026-04-30 · Category: Learning Tips

How spaced repetition really works for language learning: the forgetting curve, SM-2 algorithm, retrieval practice, plus a 20-minute daily routine.

If you have ever used Anki, Memrise, Duolingo, or any modern language app, you have used spaced repetition, even if no one told you. It is the quiet engine inside almost every memory tool of the last 30 years. The principle sounds dull, the mechanism is genuinely elegant, and once you understand why it works, you can stop fighting your memory and start working with it.

This article walks through the science from the 19th-century experiments to the modern SM-2 algorithm, explains why retrieval practice matters more than rereading, and gives you a practical 20-minute daily routine. We also cover the limits.

For a head-to-head with the other dominant language-learning method, see our piece on comprehensible input vs flashcards. For wider context, see the science behind effective language learning.

The Forgetting Curve: Where It All Started

In 1885, German psychologist Hermann Ebbinghaus published one of the founding studies of experimental memory research. His method was almost comically rigorous: he memorised lists of nonsense syllables (zof, baq, ked) and tracked how quickly he forgot them.

The pattern he found is now known as the forgetting curve:

• Within 20 minutes: you forget about 40% of newly learnt material.
• Within one hour: you have lost more than half.
• Within a day: you remember around 30 to 35%.
• Within a week: around 25%.

The hopeful discovery: each time you review and successfully recall an item, the curve flattens. Decay slows after the second exposure, slower again after the third. With a few well-timed reviews, an item you used to forget overnight becomes something you remember for years.

Modern researchers have replicated Ebbinghaus's findings many times. The spacing effect (the boost from spaced reviews compared to crammed study) is one of the most robust findings in cognitive psychology. See the Wikipedia article on the spacing effect for further reading.

From Ebbinghaus to SuperMemo: A Brief History

Ebbinghaus told us that spaced reviews work. He did not tell us exactly when to schedule them. Solving that puzzle took another century. In the 1980s, Polish researcher Piotr Wozniak, frustrated with how slowly he was learning English vocabulary, started experimenting on himself, recording thousands of review sessions and building algorithms to predict when he would forget each word.

His software became SuperMemo, and its SM-2 algorithm became the template for almost every spaced repetition tool that followed: Anki, Memrise, Quizlet's smart mode, and most language apps with built-in vocabulary review.

The principle is simple:

• Each card has an interval (when to next review) and an "ease factor" (how easy it is for you).
• After each review, you grade your recall (in Anki: Again, Hard, Good, Easy).
• A successful recall multiplies the interval by your ease factor (around 2.5 by default), so a 2-day interval becomes 5 days, then 12, then 30, and so on.
• A failure resets the interval to a short value and drops the ease factor slightly.
• Over time, easy cards drift to long intervals; difficult cards stay close.

That is the entire algorithm. Not magic. A feedback loop.

Why Spaced Repetition Works: Two Core Principles

Spaced repetition combines two well-studied effects.

1. The spacing effect

Spaced practice produces stronger memory than massed practice. Reviewing 10 words over five short sessions beats one long session of the same total time. Each time you nearly forget something and then recall it, the recall strengthens the memory trace. Massed study skips this productive struggle.

2. The testing effect (retrieval practice)

The act of recalling information from memory is itself a learning event. Reading a word over and over is far less effective than trying to recall it from a prompt. Research from the American Psychological Association shows that students who tested themselves remembered material substantially longer than students who simply reread it. Researchers call it one of the most underused findings in education.

Put the two effects together and you get spaced repetition: short, frequent recall sessions, scheduled so each item is tested just as you are about to forget it.

A Mental Model: Desirable Difficulty

Cognitive psychologist Robert Bjork coined the term "desirable difficulty" to describe study conditions that feel harder in the moment but produce stronger long-term memory. Spaced repetition engineers this on purpose. Reviewing a word the moment after you saw it feels easy and produces nothing. Reviewing it three days later feels harder and produces durable memory. A good SRS session is not effortless, and it is not supposed to be: the slight strain of "wait, what was that word again..." followed by recall is exactly the productive moment the system is designed to create.

The Tools You Can Use Today

Anki (free, open source, infinitely flexible)

Anki is the gold standard for serious spaced repetition. Free on web, desktop, and Android (paid on iOS), it gives complete control over decks, card formats, and scheduling. The learning curve is real, but the ceiling is high. Best for learners who want to build their own decks, for specific projects (medical vocabulary, kanji review), and for anyone willing to spend an hour learning the interface.

Memrise and Quizlet

Memrise wraps spaced repetition around its own course content and native-speaker videos. The free tier is decent. Best for learners who want SRS with content already prepared. Quizlet has a "learn" mode that uses similar principles: less rigorous than Anki, but easier for school-style flashcard sets.

Built-in SRS in language apps

Most modern language apps include spaced repetition under the hood. Duolingo's review feature, Babbel's review manager, and Hello Nabu's vocabulary engine all use SRS principles. The benefit: no separate tool to manage. The trade-off: less control over what comes back when. For most learners, the built-in SRS in your daily app is enough. Anki is for cases where you need more depth or material that no app covers.

A Practical 20-Minute Daily Routine

A sustainable daily SRS routine that fits into a real life:

Minute 1 to 12: review

Open your tool and clear due reviews. Aim for 100 to 200 reviews. Honesty matters: if you half-remembered a card, mark it Hard or Again. If you knew it instantly, mark Easy. The algorithm only works if you grade yourself truthfully.

Minute 13 to 18: new cards

Add 10 to 20 new cards. Ten per day creates around 100 reviews per day in steady state; 20 creates around 200. Pick a sustainable rate. A common mistake: adding 50 cards in enthusiasm, then drowning three weeks later.

Minute 19 to 20: cleanup

Suspend cards you keep failing or no longer find useful. Reformulate ambiguous cards. The deck is not a museum: it should evolve.

Twenty minutes a day, every day, will keep a vocabulary base of several thousand words alive indefinitely.

What Spaced Repetition Is Excellent For

SRS is the right tool for material that has the following properties:

• Atomic: each item is small (a word, a phrase, a fact).
• Discrete: there is a clear right answer.
• Cumulative: you need to retain it for a long time.
• Personally relevant: you have a reason to remember it.

In language learning, this means:

• Core vocabulary at any level.
• Verb conjugations (especially irregular ones).
• Kanji and other writing systems (Anki was practically built for this use case).
• Idiomatic expressions you have encountered and want to lock in.
• Specialised vocabulary for your work or interests.

For building vocabulary, spaced repetition is genuinely close to a free lunch.

What Spaced Repetition Is Not Good At

SRS has real limits, and pretending it does not produces frustrated learners.

1. Word lists are not fluency

You can review 5,000 word cards and still freeze when ordering coffee. Knowing a word in isolation is not the same as using it in real time. Fluency requires patterns, prosody, listening comprehension, and the ability to construct sentences quickly. None of that comes from flashcards alone.

2. Grammar in cards is brittle

Flashcards for grammar rules can work, but the results tend to be brittle: you remember the rule statement without being able to apply it. Grammar is better learnt through exposure and use, with a reference book for clarification. See why context is the missing ingredient.

3. SRS without input creates a strange dialect

Learners who spend most of their study time in flashcards develop oddly specific knowledge: thousands of words they cannot use, and gaps in everyday phrasing they should have absorbed naturally. Pair SRS with reading, listening, and speaking.

4. The motivation problem

Daily SRS is a discipline. Some learners thrive on it; others find it draining. If you dread your reviews every day, reduce your new card rate, change your deck, or accept that SRS is not your primary tool.

How to Pair Spaced Repetition With Other Methods

The smartest learners use SRS as one component of a broader routine:

• Daily structured app (15 to 30 minutes): introduces new vocabulary in context. Hello Nabu uses stories to teach grammar and vocabulary together.
• Daily SRS review (15 to 20 minutes): keeps that vocabulary alive.
• Daily input (10 to 30 minutes): podcasts, YouTube, articles, or graded readers. This is where grammar starts to feel intuitive.
• Weekly speaking practice (30 to 60 minutes): an exchange partner, tutor, or AI conversation tool. Where everything has to come together.

SRS is the maintenance layer. It does not produce fluency by itself; it makes sure the vocabulary you have learnt does not slip away. Hello Nabu reflects this: vocabulary is taught inside stories, then quietly reinforced through built-in spaced review. You see the same words come back in new contexts on a schedule that matches how memory actually works. See the six pillars of real fluency.

Common SRS Mistakes

A short list of pitfalls that catch beginners:

• Overloading new cards. Twenty per day is plenty. Fifty per day will drown you within a month.
• Cramming missed days. If you miss a week, do not clear all reviews in one session. Reduce new cards to zero, work through the backlog over days, then resume.
• Cards that are too long or vague. Each card should test one thing.
• Using SRS as your only method. Vocabulary in isolation is not language. Pair with input.
• Refusing to suspend bad cards. If a card has failed five times, the card is the problem.
• Skipping the grading honesty test. Marking everything Easy defeats the system.

Final Word

Spaced repetition is one of the genuinely well-evidenced techniques in modern learning science. It exploits two robust effects (spacing and testing) to push memory past the natural forgetting curve. For language learners, it is the right tool for vocabulary, conjugations, and writing systems, and the wrong tool for grammar intuition or conversational fluency.

Use it, but do not worship it. Twenty minutes a day, paired with input and speaking practice, keeps a vocabulary base alive for years. Two hours a day of pure flashcards burns you out and does not produce fluency.

Hello Nabu handles the SRS layer automatically inside lessons and roleplay practice. Vocabulary comes back when you need it, in contexts that resemble real life, completely free for individuals.

Start learning for free with Hello Nabu

Further Reading

Explore the research and tools behind spaced repetition:

• Anki: The most powerful free SRS tool
• Wikipedia: Spacing Effect: A solid overview with research links
• American Psychological Association: Research on memory and the testing effect
• Cambridge Applied Linguistics: Studies on vocabulary acquisition

Frequently Asked Questions

What is spaced repetition?

Spaced repetition is a study technique where you review information at increasing intervals over time, scheduled so that you see each item just before you would forget it. The method comes from Hermann Ebbinghaus's 1885 work on the forgetting curve and was formalised in modern algorithms like SuperMemo's SM-2. See the science of language learning for the broader context.

Does spaced repetition really work?

Yes, decades of cognitive psychology research support it. The combination of spacing (reviewing over time, not in one block) and retrieval practice (recalling rather than rereading) produces stronger and longer-lasting memory than massed study. It is especially powerful for vocabulary and facts that need precise recall. For practical use, see our guide to building vocabulary.

What is the SM-2 algorithm in Anki?

SM-2 is the original SuperMemo algorithm, also used by Anki. It schedules each card based on how easily you recalled it. A successful recall roughly doubles the interval; a failure resets it to a short interval. The system adapts to each card and each learner over time.

Is spaced repetition enough to learn a language?

No. Spaced repetition is excellent for memorising specific items (vocabulary, conjugations, kanji), but knowing words is not the same as using them. SRS works best paired with comprehensible input and active practice. Word lists alone do not produce fluency. See our breakdown of comprehensible input vs flashcards for the full comparison.

How much time should I spend on spaced repetition each day?

For most language learners, 15 to 20 minutes a day is the sweet spot. That covers around 100 to 200 reviews and 10 to 20 new cards. Beyond 30 minutes a day, returns drop sharply and burnout becomes likely. Combine SRS with daily speaking practice for the strongest results.

Related Articles

• Comprehensible Input vs Flashcards
• Story-Based Learning vs Flashcards
• The Science Behind Effective Language Learning
• How to Build Vocabulary
• Why Context Is the Missing Ingredient
• Best Free Language Learning Resources
• The Hello Nabu Difference: Six Pillars to Real Fluency
• How Long Does It Take to Learn a Language?

Start learning free with Hello Nabu