How to Build a Spaced Repetition Schedule That Sticks (Real Intervals, Not Theory)

What is a spaced repetition schedule, in one sentence?

A spaced repetition schedule is a plan for when to review each thing you're learning, spreading those reviews further and further apart over time. Instead of hammering a fact ten times tonight and never seeing it again, you see it once today, again in a couple of days, then a week later, then a few weeks after that.

The whole point is timing. You want to bump into each fact right before it slips away, because a review that catches a memory on the edge of fading does far more for long-term retention than a review when the answer is still fresh in your head. Get the timing right and a handful of reviews can lock something in for months.

The rest of this guide gives you a schedule you can start tonight, explains why the gaps expand, and shows how apps automate the whole thing so you never have to track it by hand.

What does the forgetting curve tell you about timing?

Back in 1885, Hermann Ebbinghaus memorized long lists of nonsense syllables and tested himself at intervals to chart how fast memory decays. His forgetting curve showed that newly learned material drops off sharply at first, then the loss slows and levels off. Most of the forgetting happens early, in the hours and days after you learn something.

That shape is the entire argument for spacing. If memory bled away at a steady rate, review timing wouldn't matter much. But because the drop is steepest early on, a well-placed review in the first day or two pays off enormously, and each later review can wait longer because the curve has flattened.

Ebbinghaus also noticed something that makes the whole system efficient: relearning is faster than first learning. Even when a memory feels gone, traces remain, so a quick review restores it far quicker than starting from scratch. He called this 'savings,' and it's why review sessions get shorter even as the material you're holding grows.

Why do expanding gaps beat reviewing every day?

Reviewing something daily feels diligent, but it wastes effort. When the answer is still fresh, recalling it is too easy to strengthen the memory much. The reviews that count are the ones where you have to work a little to dig the answer out.

A large meta-analysis by Cepeda and colleagues (2006), synthesizing hundreds of distributed-practice experiments, found that spacing reliably beats massing across a huge range of materials and tasks. Crucially, the best gap between reviews scales with how long you need to remember: the longer the retention interval you're aiming for, the longer the optimal gap between reviews should be.

That's the practical rule hiding inside the research. If your exam is next week, your gaps stay short. If you're building knowledge you want to keep for a year, the gaps should grow into weeks and months. Expanding intervals roughly track the forgetting curve flattening out, so each review lands at the moment it does the most good.

What's a concrete spaced repetition schedule I can follow?

Here's a simple expanding schedule that works for most material. Treat each step as the gap after a successful recall. When you nail a card, advance to the next interval. When you blank on it, send it back to the start and rebuild.

These numbers aren't magic; they're a sensible default that mirrors the forgetting curve. The principle matters more than the exact days, so adjust to your deadline and your material.

• Same day: a quick review a few hours after you first learn it, while it's still settling.
• 2 days later: the first real test, after a night of sleep has done its consolidation work.
• 1 week later: by now the easy recall is gone and you're genuinely retrieving.
• 2 weeks later: the gap widens because the memory is holding.
• 1 month later: a maintenance check that confirms it's stuck.
• Beyond: every few months, or retire the card once it's effortless.

How do you adjust the schedule for an exam deadline?

The schedule above assumes you have time. When a test is close, you compress it. Cepeda and colleagues' (2006) finding that the optimal gap scales with the retention interval cuts both ways: a short horizon means shorter gaps. For an exam in ten days, you might run same day, day 2, day 4, day 7, day 9, rather than stretching to a month.

Two habits keep a deadline schedule honest. First, front-load the hard cards; the ones you keep missing deserve the most repetitions, so don't let easy material crowd them out. Second, schedule a final pass in the day or two before the exam, even for cards that have gone quiet, because that last review is cheap insurance against the steepest part of the forgetting curve.

This is exactly what Cram automates on iPhone. You feed it your own material, typed topics, lecture notes, a textbook PDF, or a web link, and it turns that into question-and-answer flashcards in seconds, then schedules the reviews with spaced repetition and counts down to your exam date so the timing bends around your deadline. Cards come from your own source material rather than a stranger's deck, it works offline, and there are no ads, no account, and no data-selling.

How does the Leitner box system schedule reviews by hand?

If you'd rather use paper, the Leitner system gives you the expanding schedule without any math. Sebastian Leitner laid it out in his 1972 book So lernt man lernen. You set up a row of numbered boxes, each reviewed less often than the last, and let your cards move between them based on how well you recall.

A card you get right moves up to the next box, where it waits a longer interval before its next review. A card you miss drops back to box one and rejoins the frequent rotation. Over time, the cards you know drift into the slow boxes and the cards you're forgetting stay in front of you, which is precisely where your attention should go.

• Box 1: review every day.
• Box 2: review every few days.
• Box 3: review weekly.
• Box 4: review every couple of weeks.
• Box 5: review monthly, then graduate.

How do SM-2 and FSRS automate the timing for you?

Tracking boxes by hand works, but it's coarse: every card in a box gets the same interval whether it was easy or a near-miss. Software fixes that by scheduling each card individually. In the late 1980s, Piotr Wozniak built the SM-2 algorithm for SuperMemo, which computes an expanding interval per card from how hard each recall felt. Rate a card easy and the gap jumps; rate it hard and the next interval barely grows.

SM-2 is the ancestor of almost every modern scheduler, including the one in early Anki. Its successor, FSRS (the Free Spaced Repetition Scheduler), models your memory more directly. It estimates each card's current recall probability and times the review for the moment you're about to forget, aiming for a target retention rate you can set yourself.

You don't need to understand the equations to benefit. The payoff of a good scheduler is that it spends your time where it matters: more reps on the cards you're shaky on, fewer on the cards you've mastered. That's the same logic as the Leitner boxes and the same expanding-gap principle, just tuned per card and per recall.

Sources

Every claim above traces back to one of these. They're worth reading if you want the primary evidence behind the schedule.

• Ebbinghaus, H. (1885/1913). Memory: A Contribution to Experimental Psychology. The original forgetting-curve experiments: newly learned material decays rapidly at first then levels off, and relearning is faster than first learning ('savings').
• 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). A quantitative synthesis showing spaced practice reliably outperforms massed practice, and that the optimal gap between reviews grows with how long you need to remember.
• Leitner, S. (1972). So lernt man lernen. Introduced the Leitner box system, in which cards move to a higher box (a longer interval) when recalled correctly and drop back when missed.
• Wozniak, P. (SuperMemo, late 1980s). The SM-2 algorithm, which computes an expanding interval per card from recall difficulty and became the basis for most modern spaced-repetition schedulers, including the lineage that leads to FSRS.