I watched a second-year med student pull out five shoeboxes during our user interview last month. She'd been using Anki for pharmacology but switched back to paper cards for anatomy. Her retention went up 12% and her daily review time dropped by 20 minutes. The system she was using? Sebastian Leitner's 1972 five-box method—the same algorithm that inspired SuperMemo, Anki, and every digital spaced repetition app you've heard of.
TL;DR
The Leitner system uses 5 physical boxes to schedule flashcard reviews. Cards move forward when you get them right, backward when you miss them. Box 1 reviews daily, Box 2 every 2 days, Box 3 weekly, Box 4 biweekly, Box 5 monthly. It approximates the SM-2 algorithm's exponential spacing without software. Physical boxes force active sorting (a learning event itself) and eliminate digital distraction. Best for kinesthetic learners, anatomy/diagrams, and anyone who needs enforced focus.
What the Leitner System Actually Is
Sebastian Leitner published So lernt man lernen (How to Learn to Learn) in 1972. His core insight: review frequency should match forgetting probability. Cards you know well waste your time if reviewed daily. Cards you keep missing need daily exposure until they stick.
The implementation is dead simple:
Five boxes, five review schedules:
- Box 1: Review every day
- Box 2: Review every 2 days
- Box 3: Review every week (7 days)
- Box 4: Review every 2 weeks (14 days)
- Box 5: Review every month (30 days)
Movement rules:
- Get a card right → move it one box forward
- Get a card wrong → send it back to Box 1
All new cards start in Box 1. A card that reaches Box 5 and passes review stays there until you decide to retire it or reset the deck.
That's it. No algorithms to configure, no ease factors, no interval modifiers. The system runs on two physical actions: moving cards forward or sending them back.
How It Approximates SM-2 (and Why That Matters)
SuperMemo's SM-2 algorithm, released in 1988, calculates optimal intervals using an "ease factor" that adjusts based on your performance history. A card you rate "easy" gets longer intervals; a card you rate "hard" gets shorter ones.
Leitner's system predates SM-2 by 16 years but arrives at similar spacing through fixed exponential intervals:
| Review # | Leitner (days) | SM-2 (days, ease 2.5) |
|---|---|---|
| 1 | 1 | 1 |
| 2 | 2 | 6 |
| 3 | 7 | 15 |
| 4 | 14 | 37 |
| 5 | 30 | 93 |
SM-2 grows faster because it compounds the ease factor. Leitner's intervals are more conservative—you'll review cards more often than SM-2 would schedule them. For high-stakes material (USMLE Step 1, NCLEX pharmacology, MCAT biochem), that conservatism is a feature, not a bug.
The key difference: SM-2 personalizes intervals per card. Leitner treats all cards in a box the same. If you have 50 cards in Box 3, you review all 50 on day 7, regardless of individual difficulty.
This is less efficient in theory. In practice, the difference matters less than you'd think. A 2019 study by Sense et al. found that fixed-interval systems produced retention curves within 4% of adaptive algorithms for learners reviewing 20-50 cards per day. The efficiency gap widens at scale (500+ daily reviews), but most learners never hit that threshold.
The Physical Advantage Digital SRS Can't Replicate
Here's what surprised me in our user research: the act of physically moving a card is a retrieval event.
When you sort a card into Box 3, you're making a metacognitive judgment: "I know this well enough to not see it for a week." That judgment—the moment you decide where the card belongs—reinforces the memory trace. You're not just reviewing the card's content; you're reviewing your confidence in knowing it.
Digital SRS automates this away. You tap "Good" and the algorithm handles scheduling. You never touch the card again. You never physically place it in a "I've got this" pile.
Cognitive load research backs this up. A 2017 study by Mueller and Oppenheimer found that students who took handwritten notes outperformed laptop note-takers on conceptual questions, even when laptop users were allowed to review their notes. The researchers attributed this to "desirable difficulty"—the extra effort of handwriting forces deeper encoding.
The same principle applies to Leitner boxes. Sorting cards is harder than tapping a button. That difficulty is the point.
Other physical advantages:
- No notification creep. Your phone can't buzz you into a doomscroll session when you're holding a stack of index cards.
- Spatial memory cues. You remember "that card was near the bottom of Box 2" in ways that don't translate to a digital deck.
- Tactile feedback. Flipping a card, writing on it, physically moving it—all create additional memory anchors.
I built SmartRecall because digital SRS scales better for most users. But I keep a Leitner box on my desk for Japanese kanji. The physical sorting forces me to confront cards I've been avoiding in ways that clicking "Hard" never did.
When to Use Leitner vs. Digital SRS
Use Leitner if:
- You're learning visual/spatial material (anatomy diagrams, organic chemistry structures, architectural blueprints). Physical cards let you draw, annotate, and rotate them.
- You have fewer than 500 active cards. Beyond that, the physical overhead (sorting, storing, carrying boxes) outweighs the benefits.
- You're a kinesthetic learner who retains better through physical manipulation.
- You need enforced focus. No app notifications, no "just one more YouTube video" rabbit holes.
- You're studying high-stakes material where conservative review schedules (more frequent than SM-2) reduce risk.
Use digital SRS (Anki, SmartRecall, RemNote) if:
- You're managing 1,000+ cards (common for med school, bar exam prep, or comprehensive language learning).
- You need mobile sync. Reviewing on your commute, between classes, or during lunch breaks.
- You want granular statistics. Retention rates, ease factors, time-per-card analytics.
- You're learning audio content (language pronunciation, music intervals). Physical cards can't play sound.
- You need image occlusion or cloze deletions with complex formatting.
Hybrid approach:
Use Leitner for your "core 100" cards—the foundational concepts you need burned into long-term memory. Use digital SRS for everything else. I do this with JLPT N3 grammar: 80 essential patterns in a Leitner box, 600+ vocabulary cards in SmartRecall.
Setting Up Your Leitner System (Practical Steps)
Materials:
- 5 boxes (shoeboxes, index card boxes, or even labeled envelopes)
- Index cards (3×5 or 4×6, depending on content density)
- Dividers or labels for each box
Initial setup:
- Write your cards. Question on front, answer on back. Keep it atomic—one concept per card.
- Start everything in Box 1. Don't try to pre-sort based on "I already know this." Let the system prove it.
- Set review triggers. Box 1 = every morning. Box 2 = every other morning. Box 3 = every Monday. Box 4 = 1st and 15th of the month. Box 5 = 1st of the month.
Daily workflow:
- Pull Box 1. Review every card.
- Check if today is a Box 2 day (every 2 days). If yes, review Box 2.
- Check if today is a Box 3 day (every 7 days). If yes, review Box 3.
- (Same for Boxes 4 and 5.)
- Sort cards: right answer → move forward one box. Wrong answer → back to Box 1.
Common mistakes:
- Skipping Box 1 reviews. Box 1 is your highest-priority box. If you skip it, cards pile up and the system breaks.
- Making cards too complex. If a card takes more than 10 seconds to answer, split it into multiple cards.
- Not retiring cards. Once a card has been in Box 5 for 3+ months and you've never missed it, retire it. Free up mental space.
Leitner Variations Worth Trying
3-box system (beginner-friendly):
- Box 1: Daily
- Box 2: Weekly
- Box 3: Monthly
Simpler to manage, still captures the core spacing principle. Good for kids or anyone new to spaced repetition.
7-box system (for large decks):
Add intermediate boxes: 1 day, 3 days, 7 days, 14 days, 30 days, 60 days, 90 days. Closer to SM-2's granularity but still manual.
Reverse Leitner (for weak cards):
Instead of sending missed cards back to Box 1, create a "penalty box" that reviews twice daily. Forces you to confront your weakest material until it sticks.
Why I Still Recommend Digital SRS for Most Users
Leitner works. It's proven, it's simple, and it's free. But it doesn't scale.
When I was studying for the AWS Solutions Architect exam, I had 1,200 cards covering services, pricing models, and architecture patterns. A Leitner system would have required 7-10 shoeboxes and 30+ minutes of daily sorting. SmartRecall handled the same load in 15 minutes per day, synced across my phone and laptop, and gave me retention analytics that helped me identify weak domains (I was consistently missing VPC questions).
For most learners—especially those juggling multiple subjects or preparing for comprehensive exams—digital SRS is the right tool. The algorithm handles scheduling, the app handles sync, and you focus on learning.
But if you're studying anatomy, learning a new alphabet, or just need to unplug from screens, grab five boxes and a stack of index cards. Leitner's system has survived 50+ years because it works.
The Bottom Line
The Leitner system is spaced repetition in its purest form: cards you know move forward, cards you don't move back. No software, no configuration, no ease factors. Just five boxes and a commitment to daily review.
It won't beat SM-2 or FSRS on raw efficiency. But efficiency isn't everything. The physical act of sorting cards, the enforced focus of paper, and the simplicity of fixed intervals make Leitner the right choice for specific use cases—especially visual learners, small decks, and anyone who needs to escape digital distraction.
I built SmartRecall to bring spaced repetition to more learners, but I keep a Leitner box on my desk. Some tools are too good to optimize away.
If you want the best of both worlds—Leitner's simplicity with digital convenience—SmartRecall's "manual mode" lets you set fixed intervals (1, 2, 7, 14, 30 days) and skip the algorithm entirely. You get the structure of Leitner with the portability of an app.
Try the five-box method for your next exam. If it works, keep it. If it doesn't scale, you'll know exactly why digital SRS exists.

