The Forgetting Curve, and How to Get Ahead of It

Forgetting most of a lecture within a fortnight, and keeping the anecdote rather than the mechanism, is the ordinary shape of memory. It has been measured for well over a century. The part you have any control over is when you come back to the material.

What Ebbinghaus actually measured

In the 1880s Hermann Ebbinghaus ran a long series of experiments on himself. He learned lists of nonsense syllables, deliberately meaningless so that prior knowledge could not help, and then measured how much effort it took to relearn each list after a delay.

The pattern he found is the reason the forgetting curve is drawn the way it is. Loss is steep at the beginning and then flattens out. A large share of what you lose, you lose quickly. What is still there after that first drop tends to hang around considerably longer.

Two honest caveats. Nonsense syllables are not a lecture on renal physiology, and meaningful, well-connected knowledge decays more slowly than his lists did. And a curve drawn from one man's data gives you a shape rather than a timetable. What transfers is that shape: fast early loss, then a long slow tail.

Intersecting the curve

You are not going to stop the decay, so the practical question is when to interrupt it.

Each time you successfully bring something back, the curve restarts, and it restarts flatter. The next decline is slower than the last. Enough returns, spaced properly, and the tail gets long enough to cover the exam and a good while after it.

This is also why the "review everything constantly" instinct is wasteful. Reviewing something an hour after you learned it intersects the curve at the point where nothing has yet been lost, which costs time and buys you almost nothing.

The gap is the active ingredient

Take a fixed number of reviews. You can do them all in one evening or spread them across three weeks. The effort and the number of repetitions are the same either way, and the spread version produces substantially better retention.

Cepeda and colleagues pulled together a large body of distributed-practice research and found the spacing advantage holding across an enormous range of conditions: different disciplines, different ages, different kinds of thing being learned, different lengths of delay before the final test. Few findings in learning research hold up that consistently.

The reason connects to how retrieval works. A review that happens while everything is still fresh asks nothing of you, because the answer is already sitting at the front of your mind. A review that happens after you have started to forget requires you to actually go and find it, and the finding is what strengthens the memory. That is the difference between recognising something and producing it, which is the topic of why re-reading fails and active recall works.

Spacing is what makes a review worth doing, rather than an extra bit of discipline layered on top of one.

How far apart the gaps should be

Cepeda and colleagues also asked the more practical question: how long should the gap be. Their answer was that it depends on how long you need to keep the knowledge. The further away the test, the longer the optimal gap between study and review. Short horizon, short gaps. A final exam three months out justifies gaps measured in weeks.

That gives you a starting ladder rather than a rule. For a topic covered in this week's lecture, with an exam later in the term:

• a first review a day or two afterwards
• a second about a week later
• a third two or three weeks after that
• then roughly monthly until the exam approaches

Treat those numbers as an opening position, not a prescription. The adjustment rule matters more than the starting point. If a review was effortless, lengthen the next gap. If you could not produce the answer at all, shorten it. You are aiming for the gap at which you can still get there, but not without work.

Doing that adjustment by hand, per topic, across four courses, is where most people quietly give up. SymbioLearn tracks the gaps at the Concept level instead, and Today names the Concepts that deserve attention with a plain reason attached, that you have not touched something in twelve days or that it has caught you out a few times recently, and leaves the decision about what to do with it to you.

A review that felt easy was scheduled too early

Robert Bjork's work on desirable difficulties makes the case that conditions which make practice feel harder and slower often improve what you retain. Spacing is the clearest example. The uncomfortable review, the one where you sit staring at the question for ten seconds, is the one doing the work.

Two practical consequences. Do not review something the same afternoon you learned it and count that as spacing, because it is not. And do not judge a review by how smoothly it went. Smooth means you came back too soon, and the correct response is to push the next gap out further, not to feel good about it.

Being late does not reset anything

People abandon review schedules because they fall behind and conclude the plan is ruined. It is not.

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, which means you will find the gaps in your knowledge more clearly. Shorten the next gap to compensate and carry on. You do not go back to the beginning of the ladder because you missed a rung.

The only genuinely bad outcome is deciding that because the schedule slipped, there is no point reviewing at all until revision week.

Why this fits a semester better than cramming does

New lectures arrive every week whether or not you are keeping up. A rolling ladder absorbs that steadily: today you see a couple of new topics, one from last week, one from three weeks ago. The daily load stays roughly flat because old topics keep dropping out to longer and longer gaps.

Cramming is the same content arranged with every gap set to zero. It reliably produces enough performance to get through Thursday's paper and very little that survives to the following term, which is a real cost when the next course assumes you kept it.

Put the first gap in your calendar

Take the topic you covered most recently, and book two reviews now: one two days from today, one a week after that. Keep them short, and make them attempts to produce the content rather than re-read it. If you would rather see the whole ladder laid out against your actual exam date, the spaced repetition planner will build one for you.