The Forgetting Curve and How to Space Your Reviews

Most of what you forget, you forget soon after learning it. What is still there after the first few days fades much more slowly. This pattern is called the forgetting curve, and it is the reason the timing of your reviews matters: for the same amount of study, when you come back to a topic changes how much of it you keep.

What Ebbinghaus measured

The curve comes from Hermann Ebbinghaus, who tested his own memory in experiments he published in 1885. He learned lists of 13 nonsense syllables, chosen to be meaningless so that prior knowledge could not help. Then, after a delay of between 20 minutes and a month, he timed how long it took to relearn each list. The time he saved the second time round, which he called savings, measured what was left.

Twenty minutes after learning a list, relearning it saved him 58% of the original time. After an hour the saving was 44%, after a day 34%, and after a month 21%. The biggest drop came in the first hour. After the first day, the figure fell slowly.

His figures are savings in relearning time. They say nothing directly about how much of a list he could recite, so claims such as "you forget 70% of what you learn within a day" do not come from his data.

Two caveats. His numbers come from one man learning meaningless syllables, so use the shape of his curve as a guide and do not read much into the exact figures. A 2015 replication by Murre and Dros, again with one person learning lists, found a similar curve. Knowledge from real courses can show the same shape over years: when Harry Bahrick tested 733 people on Spanish learned at school up to 50 years earlier, retention fell for the first three to six years, then held steady for up to 30 years.

What a review does to the curve

Each time you successfully bring something back, the next decline tends to be slower than the last. Ebbinghaus saw this in his own data. When he relearned the same lists on six days in a row, each day's relearning took fewer repetitions than the day before. He also found that, for a 12-syllable list, 38 repetitions spread over three days did as much as 68 made on the day before.

Spreading the same number of reviews over days or weeks, instead of doing them all in one evening, is called spacing. A 2006 review by Nicholas Cepeda and colleagues gathered 839 assessments of spaced study from 317 experiments on verbal memory, such as word lists and facts. A 2013 review of ten study techniques by John Dunlosky and colleagues rated spacing as one of only two with high utility, because its benefits held for learners of different ages and abilities, on many kinds of test and in real classrooms.

One explanation: a review after you have started to forget makes you work to find the answer, and that effort seems to help. In one study, students practising Swahili–English word pairs found recall harder when the gaps between attempts were longer, and remembered more of the pairs on the final test.

How far apart to space reviews

In a 2008 study, Cepeda and colleagues taught more than 1,350 people a set of obscure facts, gave them one review after a gap of up to three and a half months, and tested them up to a year after that. The best gap, of the ones they tried, depended on how far away the test was:

Test after the review	Best gap between learning and review
1 week	1 day
5 weeks	11 days
10 weeks	3 weeks
50 weeks	3 weeks

The further away the test, the longer the best gap. A gap that was too long cost much less than one that was too short.

These were trivia facts with a single review, so treat the numbers as a guide. A course usually needs several reviews, and their exact pattern seems to matter less: in word-pair experiments by Karpicke and Roediger, making the gaps grow made little difference to long-term retention, while delaying the first test did help.

A reasonable starting schedule for a topic from this week, with the exam about ten weeks away:

1. First review: one or two days after you learn it.
2. Second review: about a week later.
3. Third review: two to three weeks after that.
4. After that, every three weeks or so until the exam.

For example, say a contract law lecture on offer and acceptance is on 6 October and the exam is on 15 December. You would review it on 8 October, 15 October, 5 November and 26 November, with a last review in the week before the exam.

How to adjust the gaps

Adjust the gaps as you go:

• If a review was easy, make the next gap longer.
• If you could not recall the answer, make the next gap shorter.

You are looking for the gap at which you can still recall the topic, but only with some effort. Elizabeth and Robert Bjork call conditions like this desirable difficulties. Spacing, mixing topics and testing yourself all make practice feel harder while often improving what you remember later. They add a condition: a difficulty only helps if you can respond to it successfully.

Two practical points:

• A second look later the same day is not much of a gap when the exam is weeks away.
• Do not judge a review by how smoothly it went. A smooth review usually means you came back too soon, so push the next gap out.

SymbioLearn keeps track of this for each Concept in a Module with a simple rule: your first correct answer sets a two-day gap, a correct answer given two or more days after you last worked on the Concept doubles the gap, up to 60 days, and a wrong answer halves it. Today then lists the Concepts worth coming back to, each with a plain reason such as "It's been 12 days since you last worked with this", and leaves the choice of what to do to you.

If you fall behind

Falling behind does not ruin a schedule. If a review was due last Tuesday and you get to it on Sunday, do it on Sunday. It will be harder than it would have been, and you will see the gaps in your knowledge more clearly.

Then set the next gap from the result: longer if you still recalled it, shorter if you did not. You do not go back to the start of the schedule because you missed a review. In Cepeda's study, a review that came too late cost much less than one that came too early.

The one thing to avoid is deciding that because the schedule slipped, there is no point reviewing until revision week.

Spacing versus cramming

A rolling schedule also copes with new lectures arriving every week: on a given day you might review a couple of new topics, one from last week and one from three weeks ago. The load stays manageable because older topics come back less and less often.

Cramming is the same content with every gap set to zero. It can hold up for a test a few days away, and it fades fast after that. In Cepeda's 2008 study, the best gap improved recall by 10% over no gap when the test was a week later, and by 111% when it was ten weeks later. That matters when a later course assumes you kept what you learned in this one.

What to do this week

1. Pick the topic you covered most recently.
2. Book two short reviews in your calendar now: one two days from today and one a week after that.
3. Make each review an attempt to recall the topic from memory. Write down what you can, then check your notes.
4. After each review, set the next gap: longer if it was easy, shorter if it was not.

To see how a topic is likely to fade and when to review it, try the forgetting curve calculator. To lay out a schedule for all your topics against your exam date, use the spaced repetition planner.