Is spaced repetition only useful for flashcards and vocabulary?

No. The spacing and testing effects apply to any material you need to retrieve later, including arguments and concepts. What changes is the prompt: for ideas you ask 'what was the claim and why', not 'what is the definition'.

How many items can I keep in rotation?

Far fewer than people expect. A handful of new ideas per week is a realistic sustainable rate. The common failure is treating capture as free and discovering the review load only after it is unmanageable.

What happens if I miss a few days?

Drop back one interval on the items you fail, and carry on. Missing days is normal; resetting everything to the start is what makes people abandon the system.

Is Anki better than a simpler schedule?

For thousands of discrete facts, yes — per-card tuning pays off at that scale. For a few dozen ideas from things you read and watch, a fixed expanding ladder performs about as well with far less overhead.

Memory & spaced repetition

Spaced repetition explained, with the actual intervals

Spaced repetition means reviewing material at growing intervals rather than all at once. Each successful recall earns a longer gap; each failure shortens it. The research finding that matters most: the best gap scales with how long you need to remember something — roughly 10–20% of your target retention window.

Asutosh·Co-founder, Recl·23 August 2026·4 min read

What is spaced repetition?

Spaced repetition is reviewing something on a schedule that expands: today, then in three days, then in a week, then in a month. Every time you successfully pull the idea out of memory, the next review moves further away. Every time you fail, it moves closer.

It combines two separate effects that both have strong evidence behind them. The spacing effect says the same total study time produces more durable memory when it is spread out. The testing effect says the act of retrieving is itself what strengthens the memory, more than re-reading does. Spaced repetition is simply the schedule that exploits both at once.

Why do gaps beat volume?

Because a memory you can retrieve effortlessly is not being strengthened. Robert Bjork's framing is that learning benefits from desirable difficulties — conditions that make performance worse in the moment and retention better later. A gap long enough to make you hesitate is one of them.

Cramming works in the opposite direction. Six reviews on Sunday night feel enormously productive because by the sixth you are fluent. That fluency is the problem: you have optimised for a state you will not be in on the day you actually need the idea.

How long should the gap actually be?

This is where most explanations wave their hands, and where the research is unusually specific.

The 2006 meta-analysis by Cepeda and colleagues, covering hundreds of experiments on distributed practice, found that the optimal gap is not a fixed number — it scales with how long you need to retain the material. As a working rule, the best spacing is somewhere around 10–20% of your target retention interval.

You need to remember it for	Roughly optimal first gap
A week	1 day
A month	3–5 days
Six months	3–4 weeks
A year or more	1–2 months

This explains a common frustration: people review daily, remember well for a fortnight, and lose it by the following term. Daily gaps are correctly sized for a one-week horizon and far too short for a one-year one.

A schedule you can actually run

If you want durable memory measured in years, and you do not want to run an algorithm in your head, this shape works:

Same day — state the idea once from memory, without the source in front of you.
3 days — recall the specific claim, not just the topic.
1 week — explain it to someone, or write it as you would send it to a friend.
1 month — connect it to something you have learned since.
3 months — a final check. If it comes back cleanly, stop scheduling it.
On failure at any step — drop back one interval rather than starting over.

That last rule matters more than it looks. Resetting to zero after one lapse is what makes people quit; a single missed recall does not undo the previous four.

What about SM-2, Leitner, and the algorithms?

They are all implementations of the same idea, differing in how they adjust the gap.

System	How it decides the next gap	Best for
Leitner (1972)	Physical boxes; a correct card moves up a box, a wrong one goes back to the first	Understanding the mechanism; still works with index cards
SM-2 (SuperMemo, 1987)	Multiplies the previous gap by an ease factor you adjust per card	Large decks of discrete facts; the basis for Anki
Fixed expanding schedule	A preset ladder like the one above	Ideas and arguments, where per-item tuning is overkill

For vocabulary or anatomy, a per-card algorithm earns its complexity. For ideas from things you read and watch it rarely does — you have tens of items, not thousands, and the difference between a 21-day and a 26-day gap is lost in the noise.

Why does spaced repetition fail for most people?

Almost never because the intervals were wrong. Three failure modes account for nearly all of it:

The deck becomes a chore. Capture is frictionless, review is not, so the backlog grows until opening the app feels like debt collection.
The prompts are bad. A card that says "Chapter 4 — incentives" cannot be retrieved, only recognised. A good prompt asks a question with a specific answer.
Nothing shows up uninvited. A system that requires you to remember to review is a system built on the exact faculty you are trying to compensate for.

The third is the fatal one. Spaced repetition is a scheduling problem with a delivery problem attached, and the delivery half is what quietly kills it.

How Recl handles this

Recl generates the prompts from the material you saved — so the question has a specific answer rather than being a filing label — and then sends reviews through Telegram or Discord on an expanding schedule. Marking one reviewed pushes the next further out; missing one brings it back sooner.

The delivery choice is deliberate: reviews arrive somewhere you already are, rather than waiting behind an app you have to remember to open.

Frequently asked questions

Is spaced repetition only useful for flashcards and vocabulary?: No. The spacing and testing effects apply to any material you need to retrieve later, including arguments and concepts. What changes is the prompt: for ideas you ask 'what was the claim and why', not 'what is the definition'.
How many items can I keep in rotation?: Far fewer than people expect. A handful of new ideas per week is a realistic sustainable rate. The common failure is treating capture as free and discovering the review load only after it is unmanageable.
What happens if I miss a few days?: Drop back one interval on the items you fail, and carry on. Missing days is normal; resetting everything to the start is what makes people abandon the system.
Is Anki better than a simpler schedule?: For thousands of discrete facts, yes — per-card tuning pays off at that scale. For a few dozen ideas from things you read and watch, a fixed expanding ladder performs about as well with far less overhead.

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).
Roediger, H. L., & Karpicke, J. D. (2006). Test-Enhanced Learning: Taking Memory Tests Improves Long-Term Retention. Psychological Science, 17(3).
Bjork, R. A., & Bjork, E. L. — work on desirable difficulties and the distinction between learning and performance.
Leitner, S. (1972). So lernt man lernen — the original box system.
Woźniak, P. (1987). The SM-2 algorithm, SuperMemo.

Asutosh

Co-founder, Recl

Builds Recl, a memory system for long-form video and reading. Writes about why people forget what they watch, and what actually fixes it.