How Spaced Repetition Algorithms Work (SM-2 and Beyond)

Every spaced repetition app looks a little different on the surface, but underneath, they're all solving the same problem: for each individual flashcard, when is the ideal moment to show it to you again? Get that timing right and you build long-term memory efficiently. Get it wrong, either too soon or too late, and you waste review time or forget the card entirely. This is what a spaced repetition algorithm actually does.

The Basic Idea Behind Every Spaced Repetition Algorithm

At its core, a spaced repetition algorithm tracks two things for each card: how well you know it, and how long it's been since your last successful review. It uses those two inputs to calculate an interval, the number of days until the card should appear again.

When you review a card and get it right, the interval grows. When you get it wrong, the interval shrinks back down, often to zero, so you see the card again almost immediately. Over many reviews, cards you consistently know well end up with intervals measured in weeks or months, while cards you struggle with stay in frequent rotation until they stick.

SM-2: The Algorithm That Started It All

Most modern spaced repetition systems trace back to SM-2, an algorithm published in 1987 by Piotr Wozniak, the creator of SuperMemo. SM-2 introduced a concept called the "ease factor," a number attached to each card that represents how easy that card is for you personally.

Here's roughly how it works. After you review a card, you rate your recall, typically on a scale from "again" (you forgot it) to "easy" (you knew it instantly). That rating adjusts the card's ease factor up or down. A higher ease factor means the interval grows faster on future reviews; a lower one means it grows more slowly, or resets if you got the card wrong.

The next interval is then calculated by multiplying the previous interval by the ease factor. A card with an ease factor of 2.5 that you last saw 10 days ago might come back in 25 days if you get it right again. This is why spaced repetition intervals grow in an accelerating curve rather than a straight line: each successful review compounds on the last.

SM-2 is the algorithm Anki is built on, and its descendants power most spaced repetition tools available today, including the scheduling logic behind Space Repeat.

Why Individual Cards Matter More Than Fixed Schedules

The key insight that makes spaced repetition work is that it operates on the card level, not the deck level. A fixed study schedule (review everything every three days, say) treats every card the same, regardless of whether you've known it for months or just got it wrong twice in a row. That's inefficient in both directions: you over-review what you already know and under-review what you don't.

A proper spaced repetition algorithm avoids this by giving each card its own trajectory. Two cards added to the same deck on the same day can end up on completely different schedules within a week, one reviewed daily because you keep missing it, the other pushed out a month because you've nailed it every time.

What Actually Changes Between Different Apps

If most spaced repetition tools use some variant of SM-2 or a similar algorithm, what actually differs between apps? Mostly it's the level of control you're given, and how much friction sits between you and your first review.

Anki exposes nearly every parameter of its algorithm: ease factor starting points, interval modifiers, how many "again" ratings before a card resets fully. That's powerful if you want to fine-tune it, but it also means new users spend real time configuring settings before reviewing a single card.

Space Repeat runs the same core scheduling logic without exposing the configuration. You create a deck, add cards, and start reviewing. The algorithm handles the ease factors and interval calculations in the background, so the scheduling works the same way it would in a more configurable tool, just without asking you to understand it first.

The Takeaway

The specific algorithm matters less than whether an app has one at all. Shuffling cards randomly, or reviewing them in the order you created them, isn't spaced repetition, no matter how the app markets itself. If you want to know whether a study tool will actually help you retain information long term, the question to ask is simple: does it track how well you know each card individually, and does it use that to decide when you'll see it again? If the answer is yes, you're looking at some form of the same algorithm that's been proven effective since 1987.