Spaced repetition: the science of remembering more, with less effort

How spaced repetition works, where it comes from, and why it let me learn 300 kanji and 1,000+ Japanese words in 18 months — with Anki for vocabulary and Kanzen for kanji.

Spaced repetition is the single most effective study habit I've found — and I use it every day. Over the past eighteen months it's how I learned 300 kanji and more than 1,000 Japanese vocabulary words, in sessions that rarely run past twenty minutes.

The idea is simple. Instead of cramming, you review each piece of information at growing intervals — ideally just before you're about to forget it. Get the timing right and a memory that would have evaporated in days can be made to last for years. Here is why that works, where the method comes from, and how to start.

Your memory fades on a predictable curve

In 1885, the German psychologist Hermann Ebbinghaus ran a now-famous experiment on himself. He memorised lists of nonsense syllables — "WID", "ZOF" — then tested how much he could recall after minutes, hours, and days. Plotting the results gave the first forgetting curve: retention drops steeply at first, then levels off. Within a single day, he had lost a large share of what he'd just learned.

100%50%0%D0D10D20D30Time since learning
The classic forgetting curve (Ebbinghaus, 1885): after one study session, recall decays fast at first, then flattens near the floor.

That same monograph also described the learning curve and the spacing effect — making Ebbinghaus's findings among the most replicated in all of psychology. The forgetting curve isn't a personal failing; it's how memory is built. The trick is to work with it.

The spacing effect: timing beats cramming

Here's the counter-intuitive part. If you have one hour to study, you'll remember far more by splitting it across several days than by spending the whole hour at once — even though the total effort is identical. This is the spacing effect, and it's one of the most robust results in learning science. A landmark 2006 review in Psychological Bulletin synthesised 317 experiments and found distributed practice reliably beat massed practice.

Each well-timed review does more than refresh the memory: it flattens the next forgetting curve, so the memory decays more slowly each time. That's why the intervals can keep growing.

100%50%0%With spaced reviewsNo review0102030Time (days)
With each spaced review (orange dots), memory returns to full strength and forgetting slows — the curve gets shallower every time. Without review (dashed), it's already gone.

Reviewing at the right moment doesn't just top up a memory — it makes the next forgetting slower. Do that a handful of times and the knowledge becomes effectively permanent.

The same research shows the ideal gap depends on your goal: the longer you need to remember something, the longer the optimal spacing between reviews. You don't have to compute that yourself — software does it for you (more on that below).

Recall, don't reread

Spacing answers when to study. The other half is how. Re-reading your notes feels productive, but it mostly builds familiarity, not memory. What actually strengthens recall is retrieval — forcing yourself to produce the answer from scratch.

In a much-cited 2006 study, Roediger and Karpicke had students either re-read a passage or test themselves on it. After five minutes, re-reading looked better. But a week later the result flipped decisively: the students who had practised retrieval remembered far more. Psychologists call this effortful-but-effective struggle a desirable difficulty — and it's exactly what a flashcard creates. Every card is a tiny test, which is why spaced repetition and flashcards are such a natural pair.

The schedule that does the heavy lifting

Put spacing and retrieval together and you get a simple loop: each time you recall a card correctly, its next review is pushed further into the future — one day, then three, a week, two weeks, a month, and so on. Cards you miss come back sooner.

Interval before the next review+1d+3d+7d+16d+35dLearnedD1D4D11D27D62
Each successful recall roughly doubles the wait until the next review. A card you know well might not resurface for months.

This is an old idea, refined over fifty years:

  • 1972 — the Leitner system. Science journalist Sebastian Leitner popularised a box method: flashcards you get right move to a box reviewed less often; cards you miss drop back. Spacing, with physical index cards.
  • 1987 — SuperMemo. In Poland, Piotr Woźniak turned this into an algorithm. His SM-2 rated how hard each card was for you and computed a personalised interval — the foundation most apps still build on.
  • 2006 — Anki. Damien Elmes wrapped a modified SM-2 into a free, open-source app (originally to learn Japanese vocabulary, fittingly). It now ships with an even smarter scheduler, FSRS, by default.

Who actually relies on this

Spaced repetition is no longer a niche trick:

  • Medical students. Anki has become a rite of passage for the USMLE licensing exams; research links student-led retrieval practice to higher scores.
  • Language learners. It's the backbone of serious vocabulary study — and the reason I could reach 1,000+ words without it feeling like a grind.
  • Duolingo. Its scheduling model, half-life regression, was trained on 13 million learning sessions (Settles and Meeder, 2016) to predict the exact moment you're about to forget a word.

How I use it every day

Two tools, one principle:

  • Vocabulary → Anki. Plain flashcards, scheduled by the algorithm, completely free. This is where most of my 1,000+ words live. If you only adopt one tool, make it this one.
  • Kanji → Kanzen. This is the app I'm building, because kanji broke the flashcard mould for me. A kanji isn't one fact but three — its meaning, its reading, and the vocabulary it appears in — and those don't stick at the same rate. So Kanzen runs three independent spaced-repetition schedules per character instead of one. Same science as Anki, adapted to the writing system.

Together they cost me about twenty minutes a day. I never decide what to study — the schedule does, and I just show up.

Start small

You don't need a perfect setup. You need a daily habit:

  1. Pick one tool and resist tweaking it. Anki for almost anything; Kanzen if you're learning kanji.
  2. Keep cards atomic — one fact per card. "What does 明 mean?" beats a paragraph.
  3. Do your reviews every day, even a short session. The schedule assumes you show up.
  4. Trust the intervals. Seeing a card less often is the system working, not a bug.

Forgetting is the default. Spaced repetition is simply the most efficient way yet found to push back against it — a few minutes today buying you years of recall.

Sources