The Leitner system
The Leitner system is a way to schedule paper flashcards with a row of boxes. A card you answer correctly moves up one box, a card you miss goes back to box 1, and higher boxes are reviewed less often.
In short
- Cards move up a box when you know them and back to box 1 when you don't. Box 1 is reviewed daily, later boxes less and less often.
- The German science writer Sebastian Leitner described the method in his 1972 book So lernt man lernen.
- It is simple and needs no software, but its intervals are fixed: every card in a box waits the same time, whatever its difficulty.
- Apps such as Anki apply the same idea with algorithms like SM-2 or FSRS, which set an interval for each card.
What is the Leitner system?
The Leitner system is a method for reviewing flashcards in which the box a card sits in decides how often you see it. All cards start in box 1. When you answer a card correctly, it moves to the next box, which you review less often; when you get it wrong, it goes back to box 1.
Cards you know drift towards the back and come up rarely, while hard cards stay at the front and come up often. That is spaced repetition done by hand: the gaps between reviews of a card grow as you learn it.
Who invented the Leitner system?
Sebastian Leitner, a German science journalist, described the system in his book So lernt man lernen (“How to learn to learn”), published by Herder in 1972.[1] He called it a Lernkartei, a learning card file.
In the version usually attributed to the book, the box had compartments of increasing size, and you reviewed a compartment when it was full rather than on set dates. Most people today use a calendar instead, with one review interval per box.
How do you set up a Leitner box?
You need a stack of flashcards, a few boxes or envelopes, and a review timetable. Five boxes is a common choice:
- Write one question per card, with the answer on the back, and put every new card in box 1.
- Decide how often each box is reviewed. The table below gives one example.
- Each study day, take the boxes that are due. Read each question and answer it from memory before you turn the card.
- Move each correct card up one box and each wrong card back to box 1. A correct card in the last box can be retired.
- Add new cards at a steady pace, not all at once, so box 1 stays manageable.
| Box | Review the box | What it holds |
|---|---|---|
| 1 | Every day | New cards and cards you just got wrong |
| 2 | Every 2 days | Cards answered correctly once |
| 3 | Every 4 days | Cards answered correctly twice in a row |
| 4 | Every 8 days | Cards answered correctly three times in a row |
| 5 | Every 16 days | Cards you know well |
This doubling schedule is a popular variant of the method, and Duolingo used a similar one when it launched.[2] To keep track, number your study days: review box 1 every day, box 2 on even days, box 3 on days divisible by 4, and so on. Because boxes follow the calendar, a card that has just moved up can come back a little sooner than its box's interval. For other timetables, see the spaced repetition schedule.
What variations of the Leitner system exist?
The method has no single official form, and three choices vary the most:
- Number of boxes. Three boxes (for example daily, every few days, weekly) suit a short course or a small deck. Five or more give longer final intervals for material you want to keep for months.
- Where a wrong card goes. The classic rule sends it back to box 1. A gentler rule moves it down one box, so one slip does not undo several correct answers; some descriptions and research models use this rule.[2][3]
- Timing. Boxes can be reviewed on fixed days, as in the table above, or when they fill up, as in the version attributed to Leitner.
Why does the Leitner system work?
The Leitner system works because it combines two of the best-supported findings about memory: spacing and retrieval. Each card is answered from memory, which is active recall, and reviews of a card are spread over days and weeks instead of crammed together.
Spacing helps with flashcards in particular. In a 2009 study in which participants learned vocabulary pairs with on-screen flashcards, spaced study beat massed study for 90% of participants, yet after the first session 72% believed massing had worked better.[4] The boxes make the spacing visible, so you don't have to trust that feeling.
What are the limits of the Leitner system?
The main limit of the Leitner system is that its intervals are fixed by the box, not fitted to each card. Three consequences follow:
- No per-card adjustment. An easy card and a hard card in box 3 wait exactly as long, so easy cards are reviewed more than they need and hard ones possibly less.
- All wrong answers count the same. A near miss and a complete blank both send a card back to box 1, and there is no grade for “correct but hard”.
- Box 1 can overflow. A 2016 model of the Leitner system predicted, and an online experiment supported, that adding new cards beyond a certain rate makes cards pile up in the first box, and fewer cards get learned overall.[3]
Duolingo, which first scheduled practice with a Leitner variant, later replaced it with a model trained on learners' data. That model predicted recall more accurately, with an average error at least 45% lower, although the Leitner method still did well on one other measure.[2]
How does the Leitner system compare with spaced repetition apps?
Spaced repetition apps follow the same principle as the Leitner system, review less often what you know well, but compute an interval for each card instead of using boxes. Anki uses SM-2, which multiplies a card's interval by an ease factor after each correct answer, or FSRS, which estimates the probability that you still remember a card and schedules the review when it falls to a target you choose.[5]
The Leitner system is still worth using when you want to study away from a screen, teach children how spacing works, or keep a small deck without installing anything. For large decks or study over several years, an app does the bookkeeping and adjusts to each card. The algorithms overview compares the main methods side by side.
Sources
- Leitner, S. (1972). So lernt man lernen: Angewandte Lernpsychologie – ein Weg zum Erfolg. Freiburg im Breisgau: Herder. Catalogue record: Deutsche Digitale Bibliothek
- Settles, B., & Meeder, B. (2016). A trainable spaced repetition model for language learning. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (pp. 1848–1858). https://doi.org/10.18653/v1/P16-1174
- Reddy, S., Labutov, I., Banerjee, S., & Joachims, T. (2016). Unbounded human learning: Optimal scheduling for spaced repetition. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (pp. 1815–1824). https://doi.org/10.1145/2939672.2939850
- Kornell, N. (2009). Optimising learning using flashcards: Spacing is more effective than cramming. Applied Cognitive Psychology, 23(9), 1297–1317. https://doi.org/10.1002/acp.1537
- Anki Manual. Deck options: FSRS. https://docs.ankiweb.net/deck-options.html