# The Forgetting Curve: How Fast Do You Forget What You Learn?

Learning · 6 min read · updated 2026-10-29

> Much of what you learn fades within a day. What Ebbinghaus’s curve really shows, what the famous 70% claim gets wrong, and how reviews flatten it.

## The short answer

The forgetting curve describes how memory fades: steeply at first, then more slowly. Ebbinghaus found in the 1880s that after a day he’d saved about a third of the effort it took to learn a list of nonsense syllables. Meaningful material fades more slowly, and each well-timed review flattens the curve, so the answer is to review before you’ve lost it, by recalling rather than rereading.

You read a chapter carefully. You understand it, and you close the book feeling that you’ve got it. A week later, someone asks you about it and you can come up with a handful of points and a vague feeling that you once knew a lot more.

That isn’t a sign of a bad memory. It’s how memory works for everyone, and it was first measured more than 130 years ago by a man who spent years memorising nonsense.

## Ebbinghaus and his nonsense syllables

In the 1880s, the German psychologist Hermann Ebbinghaus set out to measure memory itself, using himself as the subject. To stop meaning and prior knowledge from helping, he made up lists of nonsense syllables, three-letter strings like *DAX*, *BOK* and *ZIF*, and memorised them until he could recite a list perfectly.

Then he waited. After twenty minutes, an hour, a day, a week, a month, he came back and tried again. He didn’t just ask what he could remember, because the answer was often “almost nothing”. He measured how much *effort* he saved by relearning: if it took 20 repetitions the first time and 8 the second, he’d saved 60%. It’s called the savings method, and it detects traces of a memory that recall can’t.

His results are among the most famous in psychology:

| Time since learning | Effort saved when relearning |
|---|---|
| 20 minutes | about 58% |
| 1 hour | about 44% |
| 1 day | about 34% |
| 6 days | about 25% |
| 31 days | about 21% |

Plot them and you get the forgetting curve: a steep drop at first, then a long, thin tail.

*(An interactive illustration belongs here. Try it on the page: https://cognosc.com/blog/forgetting-curve)*
In 2015, Jaap Murre and Joeri Dros repeated the experiment with a modern volunteer. The curve came out close to Ebbinghaus’s, which is remarkable for a result that old. They also found a small bump at 24 hours, which may be sleep helping memory consolidate, a pattern that goes back to a 1924 study in which people forgot less over a night asleep than over the same time awake.

## What the curve does and doesn’t say

The curve gets quoted in infographics with bold claims. *You forget 50% in an hour. You forget 70% in a day.* Those figures have drifted a long way from what Ebbinghaus found, and it’s worth being clear about the distance.

**It measures nonsense syllables.** They’re about the hardest thing there is to remember, because nothing connects them to anything else. Real material, ideas you understand and facts that link to things you know, fades much more slowly.

**It was one person, learning in one way.** Ebbinghaus was a disciplined subject with a fixed method. Results shift with sleep, interest, how well you understood it, and how you studied.

**It measures savings, not recall.** The figures say how much effort was saved on relearning, not how much he could have recalled. You can forget a thing completely and still relearn it much faster than a new thing, which means the memory left a trace.

**Some memories hardly fade at all.** Harry Bahrick studied nearly 800 people who’d learned Spanish at school, testing them up to 50 years later. Most of the loss came in the first few years. After that the amount they remembered held steady for decades, in a zone Bahrick called *permastore*. How you learned it, and how often you came back to it, decided who ended up there.

The real lesson isn’t a percentage. It’s the **shape**: memory falls fastest right after you learn something and then flattens, and that shape holds for almost everything people have tested.

## How reviews change the curve

Ebbinghaus found something else, and it’s the most useful part. Each time he relearned a list, it took less effort, and after each relearning the list faded more slowly. A review does two things:

- **It brings the memory back up** to full strength, resetting the fall.
- **It slows the next fall.** Each time you recall something, it lasts longer than before.

That’s why the gaps between reviews can grow. A first review within a day or so catches the steepest part of the curve. Once the memory is a bit stronger, you can wait several days, then a couple of weeks, then a month.

*(An interactive illustration belongs here. Try it on the page: https://cognosc.com/blog/forgetting-curve)*
One more detail matters. A review helps more if the memory has faded a bit first. If you review while it’s still fresh, recall is easy and the memory barely strengthens. If you wait until you’ve half forgotten, you have to work to dig it out, and that effort is what builds it. This is the [spacing effect](/blog/spacing-effect-why-cramming-fails), and the sweet spot is just before you’d lose it.

## Review by recalling, not rereading

How you review matters as much as when. Rereading your notes feels productive but asks very little of your memory. Looking at the answer shows you the destination without exercising the path. **Testing yourself**, trying to recall before you look, makes memory stronger. That’s [retrieval practice](/blog/retrieval-practice), and the two ideas fit together: space your reviews out, and make each one a test.

## A simple schedule

For something you want to keep, such as material for an exam, a language or a skill:

1. **Learn it, and test yourself once at the end.** Try to write down the main points from memory before you leave it.
2. **Review the next day.** That’s when the fall is steepest, so a short review does a lot.
3. **Review after three or four days.** Test yourself first, then check.
4. **Review after a week or two.** If you remembered it easily, wait longer next time. If you struggled, come back sooner.
5. **Then once a month or so**, for as long as you want to keep it.

Each review can be short. Five minutes of recall beats an hour of rereading, and it’s much easier to fit in.

## What the curve means for how you learn

If you’ve ever thought *I’m bad at remembering things*, the curve offers a kinder explanation. You’re not bad at it. You’re normal. The forgetting is built in, and the way round it is to plan for it.

It also explains why learning something once and moving on so often fails, even when the first learning went well. A single pass is the start of the curve, not the end of learning. The knowledge that sticks is the knowledge that was revisited, which is a different idea from studying harder.

## Why one training day is never enough

The curve explains something every employee has lived through. A company runs a full-day course, everyone nods along, and within a month almost nothing has changed. The course wasn’t bad. It was one point on the curve, with nothing after it.

The fix isn’t a better course, it’s a follow-up: a short quiz a day later, a problem to solve a week on, a refresher a month after. The same goes for a language class that meets once a week and a driving lesson that’s a long time coming. Whenever a gap is long enough to let most of the learning drain away, you’re paying to learn the same thing twice.

## Common mistakes with the curve

- **Reviewing too early.** A review ten minutes after you learn something is mostly free recall. The gain comes from letting some of it fade first.
- **Reviewing too late.** If you leave it so long that nothing’s left, you’re relearning from scratch. The sweet spot is when it takes some effort but you can still get there.
- **Reviewing by rereading.** It feels like review, but nothing is being retrieved, so it does far less.
- **Skipping the first review.** The steepest part of the curve is the first day. One short review then is worth more than a long one later.

## Where this shows up in Cognosc

After you learn an idea in Cognosc, it checks back with one question timed to land before you’d likely forget it. Get it right and the next check comes later; miss it and it returns sooner. It’s the forgetting curve, turned into a schedule. If you’d like to see how your own habits compare with what the research recommends, try the [how learning works test](/test/learning).

## What to remember

- Forgetting is steepest soon after learning, then slows to a long tail.
- The popular “you forget 70% in a day” isn’t a law: it depends on the material and how you learned it.
- Each review brings the memory back and makes the next fade slower, so the gaps can grow.
- Review by recalling from memory, not by rereading.
- The first review matters most: do it within a day of learning.

## Questions people ask

### How fast do we forget what we learn?

Fast at first. Ebbinghaus found that nonsense syllables lost about two thirds of their staying power within a day, then declined slowly over weeks. Meaningful, well-understood material fades more slowly, and how you learned it changes the curve a lot.

### Is it true that we forget 70% of what we learn in a day?

No such figure holds for everything. It’s a loose popularisation of Ebbinghaus’s results on nonsense syllables. The real finding is the shape: rapid early loss and a slow tail. How much you lose depends on the material, how well you understood it and how you reviewed it.

### How do you flatten the forgetting curve?

Review the material after a gap and try to recall it before checking. Each successful recall makes the memory last longer. A common pattern is a review the next day, then after a few days, then a week or two later, with the gaps growing as it sticks.

### Did anyone replicate Ebbinghaus’s forgetting curve?

Yes. In 2015, Jaap Murre and Joeri Dros repeated the experiment with a modern volunteer and found a very similar curve, with a small bump around 24 hours that may be linked to sleep. It’s rare for a result from the 1880s to be confirmed this closely.

### What is the difference between forgetting and not being able to recall?

Often a memory is still there but you can’t reach it without a cue. Relearning is faster than learning from scratch even when you can’t recall anything, which shows the memory left a trace. That’s why reviewing works better than starting over.

## Sources and further reading

- Ebbinghaus, H. (1885). Über das Gedächtnis (Memory: A Contribution to Experimental Psychology).
- Murre, J. M. J. and Dros, J. (2015). Replication and analysis of Ebbinghaus’ forgetting curve. PLoS ONE, 10(7), e0120644.
- Bahrick, H. P. (1984). Semantic memory content in permastore: Fifty years of memory for Spanish learned in school. Journal of Experimental Psychology: General, 113(1), 1–29.
- Jenkins, J. G. and Dallenbach, K. M. (1924). Obliviscence during sleep and waking. American Journal of Psychology, 35(4), 605–612.
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T. and Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354–380.

---
Canonical page: https://cognosc.com/blog/forgetting-curve
From Cognosc (https://cognosc.com): free tests, articles and an AI tutor that has you reason things through.