Spaced repetition for exam revision: how it works and how to use it

You don't forget things all at once. Memory fades along a power curve: steep in the first hours and days, then flattening out ^[1]. Something you learned this afternoon might be half gone by the weekend, while something you've recalled successfully a few times over some weeks can hold for months.

Spaced repetition is the technique built on that curve. Instead of reviewing on a fixed timetable, or whenever guilt strikes, you review each item just before you'd otherwise forget it. Each successful recall flattens that item's curve further, so the next review can wait longer.

Cramming works, briefly. Massed practice genuinely lifts your performance for the next day or two, which is why it feels effective and why everyone does it. The problem is what happens after: material learned in one intense block decays much faster than the same material spread over time.

For a vocabulary quiz tomorrow, cramming tonight is a defensible choice. For an exam season where you need dozens of topics to stay solid for months, it's a treadmill: every topic you cram is quietly draining away while you cram the next one.

The key experimental result is that the best gap between reviews depends on how long you need to remember ^[2]. Short retention goals favour short gaps; long retention goals favour long gaps. Remembering something for an exam five months away calls for reviews spaced weeks apart, not the same drill every night.

For a student with a fixed exam date, that has three practical consequences:

• Starting early isn't just about total hours. Early material gets long, cheap gaps that late material can never have.
• Gaps should grow while the exam is far away, then compress as it approaches, so everything gets a final pass close enough to the day to still be fresh.
• A review scheduled after your exam date is wasted. Whatever schedules your reviews, it should know when the exam is.

You can't feel your own forgetting curve, which is why spaced repetition is usually run by an algorithm. The classic schedulers use fixed interval ladders: get a card right and the gap steps up; get it wrong and it resets. Modern schedulers such as FSRS go further and model each item's memory individually, using your history with that exact card to estimate when your chance of recalling it will dip below a threshold (typically around 90%), and scheduling the review for just before that point.

Modelling forgetting explicitly is not a gimmick. In the learning-analytics literature, adding forgetting to a mastery model closes most of the accuracy gap to far more complex approaches ^[9]. The practical upside is fewer wasted reviews: easy items drift out to long gaps, hard items come back quickly.

Spacing pairs with a second finding: relearning to criterion. When you miss a review, the fix isn't to re-read the answer and move on. It's to retrieve the item again until you produce it correctly, then space it over the following days ^[5].

The cost curve is on your side. Relearning gets dramatically cheaper each time: a concept that took real effort to learn takes moments to refresh. Skipped reviews throw away the cheapest study minutes you have.

You don't need software to use spacing. A workable manual version: keep a topic list with three columns (new, learning, solid); retrieve before you re-read; when you get a topic right, promote it and double the wait before you see it again; when you get it wrong, demote it to tomorrow. A calendar and honesty will carry you a long way.

Free flashcard tools like Anki run proper spaced repetition and have earned their reputation. The trade-off is that you build and maintain the decks yourself, and cards are only as good as what you put on them. That's a real cost in exam season, but it's a legitimate path, and the algorithm doesn't care whether the card was hand-made.

Captivate Exam Hub runs FSRS scheduling over ready-made material: real past-exam questions and flashcards mapped to your subjects' knowledge points, with review gaps that compress toward your exam date and missed items re-asked until you get them right. If you'd rather build your own system, everything above is enough to start today. If you'd rather have it run for you, there's a free trial, and the flashcard guide shows how reviews work in practice.