Key takeaways
- Forgetting is fast at first, then slows down: that's the shape of the curve Ebbinghaus measured.
- Every review done at the right moment slows the next bout of forgetting: that's spaced repetition.
- The ideal gap between two reviews depends on how long you want to remember something.
- Reviewing by testing yourself, rather than re-reading, strengthens the effect even further.
Ebbinghaus, the man who measured forgetting
Between 1879 and 1885, German psychologist Hermann Ebbinghaus ran an experiment on himself that was as simple as it was austere. He memorized lists of meaningless syllables, then measured how long it took him to relearn them after a delay. If a list took him 10 minutes the first time and 6 minutes the second, he had "saved" 40% of the work: that's the share of the memory that had survived.
His results, published in 1885, trace a curve that became famous:
What the curve really says, and what it doesn't
Two lessons still hold. First, most of the forgetting happens in the first hours. Second, the curve flattens out: what survives the first week has a good chance of sticking around a while.
But beware of the shortcuts repeated everywhere. Ebbinghaus measured time saved, not a percentage of memories. And he was learning meaningless syllables: material that carries meaning, connected to what you already know, is forgotten far more slowly. The line "we forget 70% in 24 hours" isn't a universal law.
The curve was replicated in 2015 by Jaap Murre and Joeri Dros, at the University of Amsterdam, with results very close to the original. With one interesting detail: a slight rebound around 24 hours, which the authors link to sleep.
Spaced repetition: reviewing just before you forget
Every time you review a concept, the memory strengthens, and the next bout of forgetting is slower. This is the spacing effect, one of the most solid results in psychology: for the same amount of study time, spreading your reviews over time leads to better retention than cramming it all at once. A 2006 meta-analysis covering several hundred experiments confirmed it.
Schéma illustratif : la forme est réaliste, les valeurs ne sont pas des mesures. Sans révision : R(t) = 100 × (1 + t / 0,4)^-0,5. Avec révisions aux jours 1, 3, 7 et 16 : la courbe repart de 100 et la stabilité passe à 1,5 ; 4 ; 10 ; 25 jours.
The remaining question: when to review. In 2008, a team led by Nicholas Cepeda followed more than 1,350 people to find out. The ideal gap between two reviews grows with how long you want to remember something, but more slowly than that duration.
| You want to remember it for… | Ideal gap between two reviews |
|---|---|
| A week | 20 to 40% of the delay, i.e. 1 to 3 days |
| A year | 5 to 10% of the delay, i.e. 3 to 5 weeks |
In other words: for an exam next week, review every two days. For a skill you want to keep for years, reviews every few weeks are enough, once the basics are in place.
Three methods for spacing your reviews
The Leitner system
Flashcards sorted into numbered boxes. A card answered correctly moves up a box and gets reviewed less often; a missed card goes back to the first box. Devised by German journalist Sebastian Leitner in the 1970s, it works with paper and a shoebox.
The algorithms
Software computes the ideal review date for each concept, based on your answers. SuperMemo, created by Piotr Woźniak in the late 1980s, paved the way; Anki popularized it. Effective, provided you write good cards and open the app every day.
The simple calendar
No tool? Review the next day, then three days later, then a week, then a month. It's not optimal for every concept, but it's infinitely better than cramming everything the night before.
- D+1The next dayThe most cost-effective review.
- D+3Three days laterWhat held through one night consolidates.
- D+7A week laterFragile points resurface.
- D+30A month laterWhat holds here holds for a long time.
The mistakes that cancel out your reviews
- Re-reading instead of testing yourself. Re-reading creates a sense of déjà-vu, not a solid memory. Close the course and try to retrieve it. That's the principle behind active recall.
- Reviewing everything the night before. Cramming works for the next day, barely beyond that.
- Reviewing what you already know. Focus your reviews on what's fading, not on what reassures you.
- Overloaded flashcards. One question, one short answer. A card that needs a full page to answer reviews poorly.
The right rhythm to get started
- Learn a concept, then test yourself right away without looking.
- The next day, answer two or three questions about it from memory.
- Do it again three days later, then a week later, then a month later.
- Every mistake sends the concept back to the start of the cycle.
- Add new concepts as you go: the cycle runs on its own.
Tell us what you want to learn: your course schedules the tests after each concept and spaces the reviews of your cards.
Frequently asked questions
What is the forgetting curve?
It's the curve that describes how fast we forget what we've learned. Hermann Ebbinghaus measured it in 1885: forgetting is fast in the first hours, then increasingly slow.
How long do we retain what we learn?
It mostly depends on how meaningful the material is and on review. Without review, a large part fades within days. With spaced reviews, knowledge can hold for months, even years.
When should you review to avoid forgetting?
Just before you'd forget: the next day, then at increasingly long intervals. For an exam in a week, review every one to three days; to retain over a year, reviews spaced a few weeks apart.
Does spaced repetition work for everything?
It's most effective for facts, vocabulary and concepts. For skills, like coding or solving problems, combine it with practice: redoing exercises at growing intervals.
Sources
- Ebbinghaus, H. (1885). Über das Gedächtnis. Duncker & Humblot. Traduction anglaise : Memory: A contribution to experimental psychology (1913).
- Murre, J. M. J. et Dros, J. (2015). Replication and analysis of Ebbinghaus' forgetting curve. PLOS ONE, 10(7), e0120644. doi.org/10.1371/journal.pone.0120644
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T. et Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354-380. doi.org/10.1037/0033-2909.132.3.354
- Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T. et Pashler, H. (2008). Spacing effects in learning: A temporal ridgeline of optimal retention. Psychological Science, 19(11), 1095-1102. doi.org/10.1111/j.1467-9280.2008.02209.x
- Roediger, H. L. et Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249-255. doi.org/10.1111/j.1467-9280.2006.01693.x
Passionate about teaching and new technology, I write to help everyone learn smarter.