Spaced Repetition: What It Is and Why It Works

Spaced repetition, explained

Spaced repetition means reviewing each fact at increasing intervals, timed for roughly the moment you're about to forget it. It works because two of the best-evidenced findings in learning research — the testing effect and the spacing effect — happen at the same time. Here's the mechanism, the evidence, and how to use it without wasting reviews.

The forgetting curve is the problem

In the 1880s Hermann Ebbinghaus memorised lists of nonsense syllables and measured how much he could recall over time. Retention collapsed fastest in the first day and then levelled off. Modern work has refined the shape, but the basic finding has held for well over a century: most of what you forget, you forget almost immediately.

The useful part is what happens when you review. Each successful recall flattens the curve, so the material decays more slowly than it did before. Review three or four times at the right moments and the interval stretches from days to months.

Why testing beats re-reading

The instinct after a lecture is to read the notes again. Roediger and Karpicke's 2006 experiments showed why that's a poor trade: students who spent their time being tested on material outperformed students who spent the same time re-studying it — and the re-studying group was moreconfident. Rereading produces fluency, which feels like knowledge and isn't.

Retrieval is the active ingredient. Every time you pull a fact out of memory rather than recognising it on a page, the memory gets easier to reach next time. A flashcard is just the smallest possible retrieval test.

What the evidence base actually says

Two things worth knowing if you want to judge this for yourself rather than take a study app's word for it:

Cepeda and colleagues (2006) meta-analysed well over a hundred distributed-practice studies and found spaced study reliably beat massed study, with the best gap growing longer the longer you need to remember something.
Dunlosky and colleagues (2013) reviewed ten popular study techniques for Psychological Science in the Public Interest and rated only two as high-utility: practice testing and distributed practice. Highlighting, re-reading and summarising all came out low.

Spaced flashcard review is the intersection of the two techniques that survived that review. That is the entire theoretical case for this app, and it's worth being clear that the case is for the method — the AI part only removes the typing.

How the scheduling works in FlashCard AI

You review a card and say how it went. Cards you recall confidently move to a longer interval and come back later; cards you miss come back sooner and keep coming back until they stop being missed. Each day the app surfaces only what's due, which is why a deck of six hundred cards can be maintained in fifteen minutes a day once it's mature.

Every part of this happens on your phone with no connection needed. Only generating new cards requires the internet.

The failure mode to avoid

The way people lose with spaced repetition is not bad scheduling — it's volume. Generate six enormous decks in a week and the daily due count becomes a few hundred cards, you skip a day, the backlog doubles, and you quit. Small decks, aggressively pruned, reviewed daily. The free tier's one-deck-a-day limit enforces roughly the right pace by accident.

Writing cards that are worth the review

Scheduling can't rescue a bad card. Three rules do most of the work:

One fact per card. If the answer is a list of six things, you will recall four, mark it wrong, and re-drill the two you knew. Split it.
The question must have one right answer.“Discuss the causes of X” is an essay prompt, not a card.
Test the thing that will be examined,not the thing that's easy to write. This is where AI-generated decks need your judgement — deleting a third of the cards is normal and correct.

Questions about spaced repetition

Does spaced repetition actually work for exam prep?

Yes, and it is one of the few study techniques with a strong evidence base behind it. A 2013 review by Dunlosky and colleagues in Psychological Science in the Public Interest assessed ten common study techniques and rated only two as high-utility across subjects and age groups: practice testing and distributed practice. Flashcards done on a schedule are both of those at once.

What is the forgetting curve?

Hermann Ebbinghaus, testing his own memory in the 1880s, found that recall of newly learned material drops steeply within the first day or two and then flattens. Each successful review flattens the curve further, which is why the second and third reviews are worth so much more than re-reading the material a second and third time.

How far apart should reviews be?

Expanding intervals: a day or so after learning, then a few days, then a week or two, then a month or more. The general finding from Cepeda and colleagues' 2006 meta-analysis is that the optimal gap grows with how long you need to retain the material — cramming an exam next week and building knowledge for next year need different spacing. In practice you let the scheduler handle it rather than planning it yourself.

Why is being tested better than re-reading?

Retrieving a fact from memory changes the memory — it becomes easier to retrieve next time. Roediger and Karpicke showed in 2006 that students who were tested on material outperformed students who spent the same time re-studying it, even though the re-studying group felt more confident. Feeling fluent and being able to recall are different things, and re-reading mostly buys the feeling.

How many cards should I review a day?

Fewer than you think. Twenty to forty reviews a day, done every day, will beat three hundred on Sunday. The whole mechanism depends on the schedule being kept, so the sustainable number is the correct number.

Is spaced repetition useful for understanding, or only memorisation?

Only for the recall layer — terminology, mechanisms, classifications, dates, formulas, vocabulary. It will not teach you to solve a problem or construct an argument. Use it to make the recall automatic so your working memory is free for the parts that need thinking.