The Forgetting Curve, Explained: Why You Forget and How to Stop It
TL;DR
The forgetting curve describes how memory for new material decays over time — steeply at first, then more gently. Without review, a large share of what you learn is gone within a day or two. But the curve is not fixed: every time you successfully pull something back from memory, it resets and flattens, so the next drop is slower. Spaced repetition is simply the practice of timing those recalls well, which is why a handful of short reviews can hold a fact for years.
You are forgetting this sentence already. Not because you weren't paying attention — because that is what memory does with anything new. The forgetting curve is the shape of that loss: fast at first, then slower, until what's left is either gone or permanent. The useful part is that the curve isn't fixed. Recall something at the right moment and the curve resets, flatter than before.
What the forgetting curve actually is
In 1885 the German psychologist Hermann Ebbinghaus ran one of psychology's first real memory experiments, using himself as the only subject. He memorised lists of nonsense syllables — WID, ZOF, KAJ — deliberately meaningless so that prior knowledge couldn't help him. Then he waited, and measured how much effort it took to relearn each list after twenty minutes, an hour, a day, a week.
What he found is the curve everyone now draws: retention drops sharply in the first hours, then the decline slows. The material that survives the first day or two tends to hang on much longer. Plot it and you get a steep cliff followed by a long shallow tail.
The famous percentages are probably wrong for you
You have likely seen a confident claim that people forget 50% within an hour, or 70% within a day. Those figures trace back to Ebbinghaus's self-experiment — one person, memorising nonsense. Nonsense syllables are the worst case by design: no meaning, no context, nothing to attach them to.
Material you actually understand decays more slowly. A concept that connects to something you already know has more hooks holding it in place. So treat the percentages as a shape, not a schedule. The honest version of the claim is the one that matters anyway: without review, you lose a lot, quickly, and you lose it fastest at the start.
What replicated, and what didn't
A 2015 replication of Ebbinghaus's study reproduced the overall shape of the curve — so the phenomenon is real and durable. What doesn't hold is the idea of one universal curve with fixed percentages. Your curve depends on the material, how well you understood it, and how much you already knew.
What resets the curve
Here is the part that turns a depressing chart into a method. Every time you successfully retrieve a memory, that memory gets stronger and its decay slows. The second curve is flatter than the first, the third flatter than the second. This is why five short reviews spread over a month beat five hours of study in one night — not because the total effort differs, but because each review is another reset.
| No review | Reviewed at growing intervals | |
|---|---|---|
| After 1 day | Much of it gone | Reset by the first review |
| After 1 week | Little left | Solid, needs a quick check |
| After 1 month | Effectively relearning | Still there, review is brief |
| After 6 months | Gone | Usually intact |
| Total time spent | One long session | A few minutes, spread out |
Recognition is not recall
The reset only happens if you actually pull the answer out of your head. Re-reading a highlighted page feels like reviewing, but it mostly produces recognition — the warm sense that you've seen this before — which is not the same as being able to produce it when you need it. That gap is why people walk out of an exam surprised: the material looked familiar right up until it had to be retrieved.
Testing yourself from memory is the thing that works, and it works even when it feels worse than re-reading. That's active recall, and it's the engine underneath every effective study method.
“Conditions that create difficulties for the learner can enhance long-term retention.”
Timing is the whole trick
If recalling resets the curve, when should you recall? Too soon and the answer is still sitting in working memory, so nothing is strengthened and you've wasted the repetition. Too late and it's gone, and you're relearning from scratch. The sweet spot is the moment just before you would have forgotten — hard enough to be effortful, not so hard that you fail.
That moment moves further out each time you hit it. One day, then three, then a week, then a month. Tracking it by hand across hundreds of facts is impossible, which is exactly the job spaced repetition software exists to do.
Using the curve without doing the math
- Turn what you learn into questions. A curve applies to a specific thing you can either produce or not. Vague material can't be reviewed; a question can. See how to write flashcards that actually work.
- Do the first review within a day. The cliff is steepest at the start, so the earliest review buys the most.
- Answer from memory before you check. No peeking. The effort is the point.
- Let the interval grow. Each success earns a longer gap. Each failure shortens it.
- Keep sessions short and frequent. Five minutes daily beats an hour on Sunday, because the curve doesn't care how long you sat there — only how often you reset it.
This is what knojo automates. It turns your notes into questions, tracks where every one of them sits on its own curve, and shows you only the ones about to slip — usually a five-minute stack. You show up and recall; the timing is handled.
Try it on one thing
Pick something you learned this week and write three questions about it. Answer them tomorrow from memory. That single act — one recall, at the right moment — is the entire forgetting curve intervention. Everything else is scheduling.
Frequently asked questions
What is the forgetting curve?+
The forgetting curve is the pattern by which memory for newly learned material declines over time: sharply in the first hours and days, then more gradually. It was first measured by Hermann Ebbinghaus in 1885. The curve is not fixed — each successful recall resets it and makes the next decline slower.
Do we really forget 70% of what we learn in 24 hours?+
Not necessarily. That figure comes from Ebbinghaus memorising nonsense syllables, which are deliberately meaningless and therefore the worst case. Material you genuinely understand, or that connects to what you already know, decays more slowly. The shape of the curve is reliable; the exact percentages are not universal.
How do you beat the forgetting curve?+
Review before you forget, and review by recalling rather than re-reading. Each successful retrieval resets the curve and flattens it, so the next gap can be longer. A handful of well-timed reviews — day one, day three, a week, a month — holds material far better than one long study session.
Is the forgetting curve real, or has it been debunked?+
It is real. A 2015 replication reproduced the overall shape of Ebbinghaus's original findings. What has been debunked is the idea of a single universal curve with fixed percentages that applies to everyone and every kind of material.
How soon should I review something new?+
Within a day, because the curve falls fastest at the start — an early review buys the most retention per minute spent. After that, let the gaps grow: a few days, then a week, then a month, extending each time you recall successfully.
Put this into practice with knojo.
Turn your notes into a five-minute daily review queue and keep what you learn. Launching soon on iOS and Android.
