The Leitner System: How the Box Method Works and When to Use It in 2026
TL;DR: The Leitner system is a 1972 paper method that sorts flashcards into five boxes reviewed at increasing intervals. It still works for small decks, language vocabulary, and anyone who wants to stay off a screen. For exam-scale decks of a few hundred cards or more, modern algorithms like FSRS save real time. The Leitner method is the spaced repetition gateway drug. It is not the final destination for most students.
Before there were apps, before SM-2, before anyone said the words "spaced repetition" out loud in a study group, there was a German journalist named Sebastian Leitner. In 1972 he wrote a book called So lernt man lernen — roughly, How to Learn How to Learn — and inside he described a method using nothing more sophisticated than a shoebox.
That shoebox method is the one your high school language teacher probably pulled out for vocabulary practice. It is also the conceptual ancestor of every flashcard app on the market.
I keep coming back to the Leitner system because it does one thing really well: it forces the underlying idea of spaced repetition into a shape anyone can understand in thirty seconds. No algorithm, no settings menu, no telemetry. Just boxes.
How the Leitner system works
You start with five boxes, numbered 1 through 5, and a stack of flashcards. Every new card goes into Box 1.
The rule is simple:
- Box 1: review every day
- Box 2: review every two days
- Box 3: review every four days
- Box 4: review every nine days
- Box 5: review every fourteen days (or longer)
If you get a card right during review, it graduates to the next box. If you get it wrong, no matter which box it was in, it drops all the way back to Box 1.
That demotion rule is the part most people miss. A card that slipped from Box 4 does not go to Box 3. It goes home to Box 1 and starts the climb again. The system trusts that anything you forgot was probably never as solid as you thought.
The intervals roughly double, which mirrors what later researchers like Hermann Ebbinghaus and Piotr Wozniak found about the forgetting curve. Things you have just learned slip away fast. Things you have known for two weeks decay slowly.
Why it worked in 1972 and still works today
The Leitner system was the first method to bake retrieval practice and spacing into one routine. Both effects are by now some of the most replicated findings in cognitive psychology. Retrieving information strengthens memory more than re-exposure to it does, and spacing those retrievals out works better than cramming them together.
What Leitner figured out, decades before the research consensus formed, was that you do not need a perfect schedule to get most of the benefit. You just need cards you forgot to come back sooner, and cards you remembered to come back later.
For a deck of a hundred vocabulary words, that approximation is genuinely good enough.
Where the box method breaks
The cracks show up in three places.
Big decks. Once you cross a few hundred cards, the boxes start overflowing. Box 1 is supposed to be daily, but a daily Box 1 with two hundred cards in it is not daily anymore — it is an unsustainable wall. The system has no way to relieve pressure other than for you to learn faster, which is the problem you were trying to solve.
Per-card difficulty. Every card in Box 3 gets the same four-day interval whether it is "what is the capital of France" or some obscure pharmacology mechanism that you keep almost-but-not-quite forgetting. Real cards have very different difficulty curves. Lumping them into shared intervals means the easy cards waste your time and the hard ones get demoted constantly.
Missed days. Skip three days and the entire system is out of sync. There is no graceful way to catch up. You either grind through everything that came due, treat the boxes as approximate, or restart. None of those feel good.
This is where algorithms like SM-2 and especially FSRS pull ahead. They schedule each card individually, predict the day you are most likely to forget it, and rebalance automatically when you miss days. The cost is that you have to trust a piece of software instead of looking at a box.
When to use the Leitner system anyway
Despite its limits, there are real cases where the box method still wins.
You want to stay off a screen. Phones are review tools and distraction machines in the same housing. Some students simply study better with paper, and the Leitner system is the cleanest paper method.
You are studying a small, finite set. Language vocabulary for a trip. Anatomy terms for one practical exam. Code snippets for a job interview next week. Anything under about two hundred cards fits the boxes well and does not need a per-card algorithm.
You are introducing kids or beginners to the idea. The boxes make spaced repetition visible. Watching a card physically move from Box 1 to Box 5 over a month teaches the underlying concept better than any explanation does.
You want a free, tactile method without committing to an app. Paper cards do not need batteries, accounts, or cloud sync. They survive flights, dead phones, and lost passwords.
A modern Leitner setup that actually works
If you want to try it, the original five-box rhythm is still solid. A few tweaks help:
Keep boxes physically separate. Index card boxes with dividers, an accordion folder, or even five envelopes. The friction of moving a card is part of the feedback.
Cap Box 1 at fifty cards. If new intake is filling Box 1 faster than you can promote cards out, you are adding cards faster than you can learn them. This is the same rule that breaks most app users — see how many flashcards you should review per day for the longer version.
Write short cards. The Leitner method punishes long cards harder than apps do, because every wrong answer sends you back to Box 1. Cards that test one fact at a time recover from mistakes quickly. Cards that test five facts at once become permanent Box 1 residents.
Schedule review days on the calendar. Box 2 every other day, Box 3 every Monday and Friday, Box 4 every Wednesday, Box 5 the first Sunday of each month. The original intervals are a guide, not a prescription.
Leitner versus algorithmic apps: which one belongs in your week
I will be honest: I built Imprimo because for serious exam prep, I think algorithmic scheduling wins by enough margin that the comparison is not close. Cutting daily reviews by 20 to 30 percent at the same retention level matters when you have a few hundred cards and a fixed amount of time. That is the headline number from FSRS benchmarks, and it holds up in real student decks.
But algorithms also come with overhead. Settings to tune, an opinion to trust, a screen to hold. If your study volume is small or your environment makes phones a bad idea, the Leitner system gives you most of the spacing benefit at zero software cost.
A reasonable rule:
- Under ~150 cards, paper Leitner is fine and often nicer.
- 150 to 500 cards, an app with SM-2 is a clear upgrade.
- Over 500 cards or any deck you will keep for years, FSRS-powered apps pull noticeably ahead.
The method that works is the one you will actually use this week. A perfectly tuned FSRS deck you never open is worse than a shoebox you actually flip through every morning.
Common Leitner mistakes
A few patterns I have seen students fall into repeatedly:
Adding new cards every day with no upper bound. The system does not protect you from yourself. Twenty new cards a day for a month is six hundred cards in Box 1 by the end. That is no longer Leitner. That is a paper pile.
Skipping the demotion. It feels harsh to send a Box 4 card back to Box 1 because you blanked on it once. Do it anyway. The whole accuracy of the method depends on it.
Reviewing all boxes every day. If you review Box 5 daily, the spacing collapses and you have just made yourself a very slow flashcard app. Each box has its own day.
Treating the boxes as a leaderboard. Cards in Box 5 are not "won." They are just due less often. There is no finishing the deck. Spaced repetition is a maintenance practice, not a checklist.
The honest verdict
The Leitner system is a beautifully simple introduction to spaced repetition that still works for the use case it was designed for: a person, a stack of cards, and a reasonable amount of material.
It is not what I would recommend for medical school, the bar exam, or any deck that grows past a few hundred cards. For that, the math of per-card scheduling matters too much to ignore.
If you have never tried spaced repetition before, do it with a shoebox first. Spend two weeks watching cards move between boxes. Get a feel for how spacing actually feels in your brain. Then, when the deck gets too big for the boxes — and it will — move to an app and you will know exactly what the algorithm is doing for you, because you will have done it by hand.
That is the real reason the Leitner system still matters in 2026. Not as a final tool, but as the one that makes everything else make sense.