What is spaced repetition?

Spaced repetition means reviewing material at increasing intervals timed to hit the moment you're about to forget it. Retrieving information at that edge — when recall takes effort but still succeeds — strengthens the memory far more than reviewing it while it's still fresh.

Is spaced repetition actually backed by research?

Yes — it's one of the most replicated findings in memory science. Cepeda, Pashler, Vul, Wixted and Rohrer's 2006 meta-analysis in Psychological Bulletin synthesized 839 assessments across 317 experiments, and a 2008 study of 1,354 learners found the optimal study gap improved recall by up to 111% over massed study of the same material.

How long should the gaps between reviews be?

There's no single right interval. The best gap depends on how long you need to remember the material — roughly 20–40% of a one-week horizon, shrinking to about 5–10% for a one-year horizon (Cepeda et al., 2008). Because the ideal gap differs per item and grows over time, algorithms like FSRS adjust each card individually.

Why is cramming worse than spacing it out?

Cramming builds short-term memory that feels like mastery but fades fast, because long- and short-term memory form through different processes. A student who crammed a month ago has forgotten nearly everything; a student who spaced their reviews retains most of it — and each later review takes less time.

How Spaced Repetition Works (And Why You're Probably Doing It Wrong)

Q: Is expanding the interval better than fixed spacing?

Not necessarily. Karpicke and Roediger (2007) found expanding intervals helped short-term recall, but equally-spaced reviews produced better long-term retention. What matters most is making the first review hard enough by delaying it — not gradually stretching later reviews.

i have adhd, and for years “use spaced repetition” was advice i nodded at and never actually followed. i’d install anki, build a deck, review it twice, and never open it again. i understood the theory completely — i just couldn’t run the schedule by hand. that gap, between knowing the principle and actually doing it, is most of why i ended up building study date.

Spaced repetition is one of those ideas that everyone nods at and almost nobody does correctly. The advice sounds obvious — review material at intervals, don’t cram — but the mechanism behind it is less obvious, and missing the mechanism means missing most of the benefit.

What your brain is actually doing

Ebbinghaus mapped the forgetting curve in 1885. The shape is brutal: within a day of learning something, you’ve forgotten most of it. Within a week, most of what remains is gone. This isn’t a flaw — it’s your brain being efficient, clearing out information it doesn’t think you need.

The way to override this is to convince your brain that you actually need the information. And the signal your brain responds to is retrieval — specifically, effortful retrieval.

Every time you successfully recall something, the memory gets stronger and the next forgetting curve is shallower. But here’s the part people miss: the retrieval has to be effortful. Reviewing something 30 seconds after you learned it is so easy that almost nothing is reinforced. The benefit comes from retrieving it right at the edge of forgetting — when you almost can’t remember it, but you still can.

Spaced repetition is the practice of scheduling reviews to hit that window consistently.

What the research shows

The spacing effect isn’t a productivity-blog theory — it’s one of the most robust findings in memory research. A meta-analysis by Cepeda, Pashler, Vul, Wixted and Rohrer (2006), published in Psychological Bulletin, synthesized 839 assessments of distributed practice across 317 experiments. The verdict was unambiguous: spreading study out beats massing it together, across a wide range of material and learners.

The size of the effect is the surprising part. In a study of 1,354 participants, Cepeda and colleagues (2008) had people learn obscure facts with study gaps ranging from zero to 105 days. For the longer retention intervals, the optimal gap improved recall by 59% to 111% compared with studying the same material back-to-back — identical total study time, dramatically different results.

That 2008 study also answered the question everyone asks: how long should the gap be? There’s no fixed answer. The optimal gap scales with how long you need to remember something — roughly 20–40% of a one-week horizon, shrinking to about 5–10% of a one-year horizon. The interval that’s perfect for a quiz next Friday is wrong for a board exam next spring.

One popular piece of advice deserves a caveat. The idea that you should always expand the gap between reviews is overstated: Karpicke and Roediger (2007) found expanding intervals helped short-term recall, but equally-spaced reviews produced better long-term retention. What mattered most wasn’t stretching the later reviews — it was making the first review hard enough by delaying it. A good algorithm handles that nuance; a hand-drawn calendar almost never does.

And the forgetting curve underneath all of this is real. Murre and Dros (2015) successfully replicated Hermann Ebbinghaus’ 1885 forgetting curve more than a century later — the steep early drop-off is a genuine, measurable property of memory, not a Victorian curiosity.

The mistake most people make

A lot of students hear “space out your studying” and create a calendar: review Chapter 3 on Monday, again on Wednesday, again next Monday. That’s better than cramming. It’s not spaced repetition.

The problem is that the right interval is different for every piece of information and every person. Something you find easy should have a longer gap before the next review. Something you keep forgetting should come back sooner. A fixed calendar can’t account for this — it treats everything the same, which wastes time on material you already know and under-serves the things you keep forgetting.

This is what an algorithm does that a calendar can’t. Software like Anki or Study Date tracks how you performed on each card and adjusts the next review interval accordingly. Easy recall means a longer gap. Hard recall means a shorter one. The current state of the art, FSRS, goes further than older schemes like SuperMemo’s SM-2: instead of a single “ease” number, it models each memory with three variables — difficulty, stability, and retrievability — and schedules the next review to catch you right before retrievability drops too far.

this is the exact point where i kept falling off. anki is genuinely brilliant, and i still bounced off it more than once — not because the algorithm was wrong, but because doing it properly demands a level of daily setup and discipline my brain doesn’t reliably produce. and most people i watched use it made the same quiet mistake i did: they treated the review queue as optional and skipped the cards that felt hard, which turns a retrieval system back into a re-reading one.

Why cramming feels like it works

The night before an exam, cramming produces real short-term retention. You go into the exam feeling like you know the material, and often you do — for the next 24 hours.

The problem is that short-term and long-term memory formation are different processes. Cramming fills the former without doing much for the latter. A student who cramming-studied a month ago has forgotten nearly everything. A student who spaced their reviews has retained most of it — and each subsequent review takes less time.

The compounding effect is significant. Spaced repetition is slow to start and fast later. Cramming is fast to start and gone later. If you’re studying to pass a single test and never think about the subject again, cramming might be the rational choice. If you’re studying to actually build knowledge — for a profession, for a degree that builds on itself — cramming is a trap.

my honest take: cramming is a perfectly rational tool for the throwaway exam — the requirement you’ll pass once and never think about again. the trap is using it for anything that compounds. i crammed my way through plenty of those one-off exams with no regrets. the courses i actually needed later are the ones i wish i’d spaced.

How to start without building a system

The research can make spaced repetition sound like it needs a spreadsheet and a PhD. It doesn’t. The whole thing collapses to one habit: review at increasing gaps, and let your performance set the next gap.

In practice that means using a tool that schedules for you — Anki, or Study Date — rather than a paper calendar that treats every card the same. Make a handful of cards from today’s material (one idea each, phrased as a question), and review whatever the app puts in front of you each day. That’s it. The algorithm stretches the gap on cards you know and shrinks it on the ones you don’t — which is exactly the part a human can’t eyeball.

And pair it with active recall: spaced repetition decides when to review; active recall decides how. The “when” without the “how” is just a well-timed re-read — and re-reading, as it turns out, barely works.

Sources

Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354–380.
Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. (2008). Spacing effects in learning: A temporal ridgeline of optimal retention. Psychological Science, 19(11), 1095–1102.
Karpicke, J. D., & Roediger, H. L. (2007). Expanding retrieval practice promotes short-term retention, but equally spaced retrieval enhances long-term retention. Journal of Experimental Psychology: Learning, Memory, and Cognition, 33(4), 704–719.
Murre, J. M. J., & Dros, J. (2015). Replication and analysis of Ebbinghaus’ forgetting curve. PLOS ONE, 10(7), e0120644.
Open Spaced Repetition. Free Spaced Repetition Scheduler (FSRS) — scheduling built on the DSR (difficulty, stability, retrievability) memory model.

Study Date uses the FSRS algorithm — the current state of the art in spaced repetition research — to schedule your reviews. You rate how well you remembered each card, and the algorithm handles the rest. Get started free.