A student sits down with a stack of flashcards for AP Bio. They flip through the same card — "What is the function of the mitochondria?" — eighty times in a week. They still blank on it during the practice exam.
The cards aren't broken. The schedule is. Reviewing the same fact over and over in one sitting feels like studying, but the brain barely encodes it. Spaced repetition fixes that, and once you understand why, the whole approach to flashcard study changes.
The forgetting curve is real
Hermann Ebbinghaus ran the original experiments on himself in the 1880s. He memorized lists of nonsense syllables, then tested how much he could recall over time. The pattern was sharp: memory drops fast in the first 24 hours, then slower over the following days.
Modern replications show the same shape. Without review, you forget roughly half of new material within a day or two. Most students treat this as a personal failing. It's not — it's how memory works by default.
The fix isn't more review. The fix is review timed to the moment right before you forget.
Why spacing beats massing
When you study the same material in one long block, you get a brief illusion of mastery. Karpicke and Roediger's 2008 study showed that students were terrible at predicting what they'd actually remember a week later — the in-the-moment confidence didn't translate.
Cepeda et al. ran a 2008 experiment that nailed down the rule. They had students learn facts and tested them later, varying the gap between study sessions. The optimal gap turned out to be proportional to how long you want to remember the material. If you want to remember something for a week, space your reviews about a day apart. For a month, space them about a week apart.
This is the spacing effect, and it's one of the most reliable findings in cognitive psychology. Bjork calls it a "desirable difficulty" — the review feels harder because you've started to forget, but that small effort to retrieve is exactly what cements the memory.
SM-2 was the first attempt
Piotr Wozniak built the first practical spaced repetition algorithm in 1985 for SuperMemo. The version most people know is SM-2, which is what Anki still uses. The idea: after each review, you rate how easy the card was. Easy cards get longer intervals. Hard cards reset.
SM-2 works. It's also showing its age. The algorithm assumes a fixed difficulty curve and doesn't adapt well to individual memory patterns. It also schedules forever — the default assumption is that you want to remember everything indefinitely, which is wrong if you have a test in three weeks.
FSRS is the modern upgrade
The Free Spaced Repetition Scheduler — FSRS — was developed by the open-source spaced repetition community and refined through several published iterations. The current FSRS-6 model uses three latent variables for each card: difficulty, stability, and retrievability.
Retrievability is the probability you'll recall the card right now. The formula behind it is R(t) = (1 + factor × t / S) ^ (-power), where t is days since last review, S is stability, and the constants come from fitting real review data. You don't need to memorize that. What matters: FSRS predicts when your recall probability will drop to a specific target — 90% by default — and schedules the next review at that moment.
The result is fewer reviews for the same retention. Studies on FSRS-5 from the open-source team showed roughly 20–30% fewer required reviews compared to SM-2 at matched accuracy. That gap matters when you have 800 cards and four hours a week.
The exam-date problem
Classic spaced repetition optimizes for indefinite retention. That's the wrong target if your exam is May 12.
If you have AP Calc on May 12 and you start studying in early April, an SM-2-style schedule will keep stretching intervals: cards you nailed will come back in October. That's fine if your goal is to remember calculus for life. It's not fine if your goal is to peak on test day.
An exam-date-aware schedule does the opposite. It treats your test date as a hard deadline and compresses intervals as the date approaches. Cards land at high recall probability on May 12, not on May 30. StudyLess uses your exam date as a parameter in the FSRS scheduler, which is the single biggest difference between getting a 5 and getting a 3 on a flashcard-friendly AP exam. We dig into setting the right exam date in a follow-up post.
What this looks like in practice
A new card comes in at 0% retrievability. You rate it Good. The algorithm gives it a stability of, say, 3 days. Three days later your retrievability has dropped to ~90% and the card surfaces. Nail it and stability jumps to maybe 8 days; miss it and stability drops back to 1 day. Repeat that for 600 cards across six weeks of AP Bio prep, and what comes out the other side is a deck where every card is timed to maintain ~90% recall on your exam date — without you scheduling any of it.
When flashcards aren't enough
Spaced repetition is great for facts: vocabulary, formulas, dates, definitions. It's weaker for understanding. If you keep failing the same card, the issue is usually that you don't get the underlying concept — and no scheduling tweak fixes that.
That's the boundary where you switch to study guides or worked examples. Flashcards drill recall on top of understanding. They don't build the understanding itself. We cover the difference more in active recall vs. re-reading, since both topics share the same underlying mechanism.
The takeaway
Memory is leaky by default. Without spaced review, you'll forget most of what you learn within a few days. The spacing effect — confirmed across decades of research, from Ebbinghaus through Cepeda — gives you a way to fight that.
What's changed in the last few years is the algorithm. SM-2 was a great 1985 invention. FSRS, calibrated against millions of real reviews and aware of your specific exam date, is what 2026 spaced repetition should look like. The math runs in the background. Your job is just to open the app and review the cards it shows you.