Spaced Repetition Explained: From the Leitner Box to FSRS

If you adopt one study technique this year, make it spaced repetition. Nothing else does as much to help you remember more while studying less. And once it clicks, the rest of good flashcard practice tends to fall into place. (For the wider picture, start with how to study with flashcards.)

The problem: the forgetting curve

In the 1880s, Hermann Ebbinghaus tested his own memory on lists of nonsense syllables and tracked how fast they faded. The shape he found, the forgetting curve, is steep. Without review you lose a big chunk of new material within a day, and most of it within a week.

Every student knows this in their gut. The usual response is to repeat by volume: read it again, and again, and again, all in one sitting. That works for about a day, then collapses. Massed practice fights the forgetting curve, and it loses.

The fix: gaps that grow

Spaced repetition flips the approach. Instead of piling reviews together, you review an item a handful of times with the gap widening each round. Maybe one day, then three, then a week, then three weeks, then a couple of months.

Two things make it work:

• The spacing effect. The same number of reviews retains far more when they're spread out than when they're crammed. It's one of the most replicated results in memory research; Cepeda's meta-analysis is a good entry point.
• Desirable difficulty. Reviewing just before you'd forget makes recall a little effortful, and that effort is what fixes the memory. Too early and you waste the rep. Too late and you've already lost it.

The whole trick is timing each review to land near the bottom of the curve, just before it hits zero. Get that right and a memory that would have vanished in days can hold for months after only a few reviews.

The Leitner box: spaced repetition on paper

The first practical system was Sebastian Leitner's box, from 1972. Picture a row of compartments. New and failed cards sit in box 1 and come up often. Get a card right and it moves to box 2, which you review less often. Right again, box 3. Miss one anywhere and it drops straight back to box 1.

It's clever, and it works. But it has a ceiling. Once you've got a few hundred cards spread across the boxes, working out what's actually due today becomes a chore, and the intervals are only ever as fine as the boxes you built. That's the gap software fills.

The algorithms: SM-2 and FSRS

Software swaps the physical boxes for a per-card schedule. You review a card, rate how it went, and the algorithm sets the next date.

SM-2

SM-2 came out of SuperMemo in 1987, and it's still the default in a lot of apps, including StudyCards today. Every card carries an "ease factor." Rate a card well and its interval stretches; rate it badly and the interval shrinks and the card comes back soon. Simple, transparent, and good enough that it carried serious learners for decades.

Its weakness is that the formula is fixed. It doesn't really model how you forget, so it tends to over-review your easy cards and under-review the hard ones.

FSRS

FSRS, the Free Spaced Repetition Scheduler, is the modern answer (it took off from 2022 onward). Instead of one fixed formula, it fits a small model of memory, estimates your odds of recalling each card over time, and schedules reviews to hit a retention target you pick, say 90%. In practice that usually means the same retention for noticeably fewer reviews.

Think of it this way. SM-2 asks how much to grow an interval. FSRS asks when you're actually about to forget, and how confident it is in that guess.

StudyCards runs SM-2 today and is moving to FSRS as the product matures. For you as a learner the habit is identical: rate honestly and let the scheduler pick the day.

Making it actually work

The algorithm is half the system. You're the other half.

• Do what's due, and stop there. The point of spacing is not grinding everything daily. Trust the queue.
• Rate honestly. "Got there after a few seconds" is not "Easy." Generous ratings stretch intervals too far and you forget; harsh ratings waste reps. Aim for accurate.
• Keep cards atomic. No schedule can rescue a card that crams five facts together. See how to make good flashcards.
• Don't ghost the deck for two weeks then binge. A short daily session keeps the queue small. Skip a fortnight and it balloons into a wall.

Why it's worth the habit

Spaced repetition is the reason students facing huge volumes (languages, medicine, law) lean on it so heavily. It quietly turns "study harder" into "study at the right moment," and the right moment is exactly the thing software is good at working out.

Pair it with decent cards and honest ratings and you get what the research promises: durable memory for a fraction of the hours.

Want to hand the scheduling to an algorithm? Start a free deck and add your first cards in a few minutes.