Spaced Repetition for Language Learning: The Complete Method

TL;DR: Spaced repetition is the only evidence-backed method for durable vocabulary retention in language learning. With 20 cards per day and the FSRS-5 algorithm, you can reach 1,000 memorized words per month with 25% fewer reviews than classic flashcards. The key: separate active and passive cards from day one.

1. Why spaced repetition beats vocabulary lists

In 1885, Hermann Ebbinghaus demonstrated something brutal: without review, you forget 50% of what you learned within 24 hours. 80% within a week. The forgetting curve is not a metaphor — it is a physiological reality tied to how memory traces consolidate.

Vocabulary lists fail for one specific reason: they ignore this curve. You read "Fernweh = longing for distant travel" twenty times in one evening. By the next morning, it is gone.

Spaced repetition does the opposite. It schedules review exactly before you forget — not too early (wasted effort), not too late (relearning from scratch). Each successful review pushes the interval further: 1 day, 3 days, 10 days, 1 month, 3 months. Karpicke and Roediger (2008) confirmed that this retrieval practice mechanism — actively recalling information — is significantly more effective than passive re-reading for long-term retention.

The concrete result: with 30 minutes per day, a serious learner can maintain several thousand words in active memory. With lists, half disappear before the exam.

2. Active vs passive vocabulary: two different card types

This is the mistake 90% of language learners make. They create a card in one direction, review in one direction, and hope it works both ways. It does not.

Recognizing "Fernweh" when you read it in German (recognition) and using it spontaneously in a sentence (production) call on distinct neural circuits. Both must be trained separately.

The rule is simple: for every important vocabulary word, create two cards — one in each direction. The volume doubles, but so does coverage.

For learners targeting a certification (TOEFL, IELTS, DELF, Goethe-Zertifikat), production cards are non-negotiable. The speaking and writing sections do not show you the word — they ask you to generate it. If you only reviewed in the recognition direction, you will freeze.

One more thing about context. Never create a card with a bare word, no sentence. "Break" without context is useless. "It is make-or-break for the project" or "I need a coffee break" — there, the word gains meaning, attaches to a mental image, and sticks.

3. Conjugations and idioms: the traps to avoid

Conjugations

Many learners think that understanding a conjugation rule is enough to remember it. Wrong. Understanding is necessary; producing automatically is something else entirely.

In Spanish, you may understand the present subjunctive for three weeks. But mid-sentence, under pressure, you produce "habla" instead of "hable". Why? Because automatic production demands hundreds of exposures, not a grammar explanation.

Production cards are mandatory for conjugations. Recommended format:

• Front: "Conjugate 'hablar' in the present subjunctive, 3rd person singular (Spanish)"
• Back: "hable — Quiero que él hable más despacio"

The example sentence is not decorative. It anchors the form in real usage — the brain retains syntactic patterns, not abstract paradigms.

For morphologically rich languages (German, Russian, Latin), create cards for each grammatical case of high-frequency words. Yes, that is a lot of cards. No, you cannot skip it.

Idioms

Context is not optional here — it is structural. "Spill the beans", "kick the bucket", "let the cat out of the bag" — these expressions make no sense if you memorize them word by word.

Required format for an idiom card:

• Front: the situation or meaning (not the expression itself)
• Back: the full expression + sentence in context + register (informal/formal/slang)

Concrete example:

• Front: "How do you say in English that you reveal an interesting secret? (informal)"
• Back: "spill the tea — 'She spilled the tea about the breakup' (= to reveal gossip or information)"

4. FSRS-5 vs SM-2: which algorithm to use?

Anki, the most widely used flashcard tool, has run on the SM-2 algorithm developed in 1987 by Piotr Wozniak for decades. SM-2 was revolutionary for its time. Today, it is outdated.

FSRS-5 (Free Spaced Repetition Scheduler, version 5) is an open-source algorithm published in 2022, trained on tens of millions of real reviews. The difference is not cosmetic.

The Expertium benchmark (2024) compares both on real-world data:

| Criterion | SM-2 | FSRS-5 | |---|---|---| | Target retention rate | 85% | 85% | | Reviews/day for 10,000 cards | ~100 | ~75 | | Forgetting prediction accuracy | Moderate | High | | Adaptation to individual profile | No | Yes (via optimization) |

25% fewer reviews per day for the same retention level. Over a year of active study, that is 30 to 50 hours saved. Not negligible when you are preparing for a standardized test.

FSRS-5's advantage comes from two things. First, it models memory with a three-component model (stability, difficulty, retrievability) instead of SM-2's single parameter. Second, it optimizes on each user's review history — parameters adjust to how you memorize, not to a population average.

For language learning, this personal calibration is especially valuable: some learners pick up vocabulary quickly but forget conjugations fast. FSRS-5 detects this and calibrates intervals accordingly.

5. The 1,000-words-per-month strategy: a concrete plan

1,000 words in 30 days works out to 33 words per day. With one active + one passive card per word, that is 66 new cards daily. Too aggressive for most people — reviews pile up quickly and you end up overwhelmed by week two.

The realistic plan for 1,000 words per month:

• 20 new words per day (40 cards: 20 active + 20 passive)
• ~25 minutes of daily review (new cards + scheduled reviews)
• Absolute priority on high-frequency words: for English, the Oxford Academic Word List of 2,000 words covers 90% of academic texts. Start there.
• Group by semantic field: memorizing 20 words on the theme "economics" the same day strengthens associations and reduces confusion.

A few non-negotiable rules:

Never put more than 3 pieces of information on a card. Word, translation, example sentence. That is it. If you add pronunciation, etymology, and two synonyms, the card becomes a study sheet — ineffective for recall.

Create your own cards, do not import everything. Pre-built decks exist and save time. But the 20 words you personally extracted from an article you were reading stick 3x longer than words copied from an anonymous list. Active encoding during creation amplifies retention.

The 80% success rule. If you pass more than 90% of your daily reviews, increase new cards. If you drop below 70%, stop adding new cards — clear the backlog first.

Two resources to populate your decks quickly:

• For English: Diane's PDF to flashcards tool automatically extracts vocabulary from your course materials and articles.
• For any language: start with frequency lists (the top 2,000 most-used words) available on Wiktionary, then add vocabulary specific to your domain (medicine, law, science, business).

To go deeper on the underlying mechanics, the spaced repetition page covers the full theory, and the FSRS algorithm page explains the technical parameters. Active recall is the essential companion method — retrieval practice is what makes spaced repetition work.

6. Frequently asked questions