I have a bad memory. And I am ambitious. So I’ve always had to figure out ways to stay organized, learn and do things in ways that suit my particular set of strengths and weaknesses, so that I can

Most people study everything with equal frequency. They flip through the same stack of flashcards, re-read the same notes, and spend as much time on material they already know as on material they keep getting wrong. This feels productive. It is not. In 1972, a German science journalist named Sebastian Leitner published a system that fixes this problem with nothing more than a set of flashcards and a few boxes.

Five Boxes, One Rule

The system works like this. You have five boxes, each with a different review schedule. Box 1 is reviewed every day. Box 2 every two days. Box 3 every four days. Box 4 every week. Box 5 every two weeks.

All new cards start in Box 1. When you review a card and get it right, it moves to the next box. When you get it wrong, it goes back to Box 1, no matter where it was. That is the entire system.

What emerges is a self-sorting mechanism. Cards you struggle with pile up in Box 1, where they get daily attention. Cards you know well graduate to later boxes, where they take up almost none of your time. Your effort automatically concentrates on your weakest points. You stop wasting time reviewing things you already know.

Why It Works

Leitner published his system before most of the cognitive science that validates it existed. But his boxes implement three well-established principles of memory.

The first is spaced repetition. Reviewing material at increasing intervals produces stronger long-term retention than cramming. Hermann Ebbinghaus documented this in 1885, and every subsequent study has confirmed it. The Leitner boxes implement spacing mechanically: each promotion roughly doubles the interval before the next review.

The second is active recall. Looking at a flashcard and attempting to retrieve the answer from memory is more effective than passively re-reading. Every card review is a retrieval attempt, which strengthens the memory trace whether the attempt succeeds or fails.

The third is desirable difficulty. Memory consolidation is strongest when retrieval is effortful but successful. By spacing reviews so that each one happens before you would forget, the system keeps retrieval in the productive difficulty zone. Cards that are too easy get pushed to longer intervals. Cards that are too hard stay at short intervals until they stick.

The Details That Matter

Card design makes or breaks the system. Each card should test one discrete fact. Cards that bundle multiple concepts create ambiguous grading: if you know three of four items on a card, did you get it right or wrong? Atomic cards produce cleaner sorting and more reliable promotion decisions.

The specific intervals are guidelines, not laws. Leitner’s original schedule is a reasonable starting point, but the core principle is that each successive box has a longer interval than the one before it. The progression matters more than the exact numbers.

The failure penalty is deliberate. Sending a failed card all the way back to Box 1, rather than dropping it one box, reflects a real property of memory: a forgotten card needs to be relearned, not reviewed slightly more often. Some modern implementations soften this. Both approaches work, but the original rule converges faster.

Daily workload stays roughly constant even as your total card count grows. Most of each session is spent on Box 1. Later boxes contribute cards only on their scheduled days. This is one of the system’s underappreciated strengths: it scales gracefully.

Leitner vs. Algorithms

Digital spaced repetition systems like Anki and SuperMemo extend Leitner’s core idea with continuous scheduling algorithms. Instead of five discrete boxes, they calculate an optimal review date for each individual card based on its full history of correct and incorrect answers. Each card gets its own “ease factor” that adjusts with every review.

This is better for large card sets. The Leitner system assumes all cards in a given box need approximately the same review interval, which is not always true. Algorithmic systems handle this heterogeneity well.

But the Leitner system has three advantages that software cannot replicate. It requires no technology. A set of index cards and a few labeled sections in a box are sufficient. The discrete box model is easier to understand and explain than continuous scheduling algorithms, which makes it easier to start using. And for moderate-sized card sets (under a few hundred), the efficiency difference between Leitner boxes and algorithmic scheduling is small. The biggest gain comes from doing spaced repetition at all, not from optimizing the specific interval curve.

What It Does Not Do

The system optimizes for retention of discrete facts. It does not address deeper understanding, application, or transfer of knowledge to new contexts. Knowing that mitochondria are the powerhouse of the cell is not the same as understanding cellular respiration. The Leitner system is a memorization tool, and it works best when paired with other forms of study that build comprehension.

It also requires discipline. Skipping a scheduled review day means those cards are overdue, and the backlog compounds. The system is simple, but simple is not the same as easy.

References

  • Sebastian Leitner, So lernt man lernen (1972)
  • Hermann Ebbinghaus, Memory: A Contribution to Experimental Psychology (1885)