The Ebbinghaus Forgetting Curve: What It Really Shows

The forgetting curve describes how memory for newly learned material fades over time when you do nothing to reinforce it. Plotted on a graph, it drops steeply at first — most of the loss happens soon after learning — and then flattens out, so what survives the first day or two tends to stick around far longer.

That basic shape is the durable part of Ebbinghaus's work: forgetting is fast early and slow late. The exact numbers behind the curve are where care is needed, because what he actually measured isn't quite what the popular graphs imply.

Hermann Ebbinghaus (1850–1909) wanted to study memory scientifically at a time when most people thought that was impossible — memory seemed too tangled up with meaning, interest, and prior knowledge to measure cleanly. His solution was to strip meaning out entirely.

He invented the nonsense syllable: a consonant-vowel-consonant unit like WID or ZOF that carried no prior meaning. By memorising long lists of these, he could measure pure rote learning without the advantage of already knowing what the material was about. And he ran the experiments on a single subject he could fully control — himself, learning and relearning lists, over and over, for years.

He published the results in 1885 as 'Über das Gedächtnis,' later translated into English as 'Memory: A Contribution to Experimental Psychology.' It effectively founded the experimental study of memory.

Here's the detail almost every blog post gets wrong. Ebbinghaus didn't measure 'what percent can I still recall.' He measured savings: how much less effort it took to relearn a list the second time. He'd learn a list to the point where he could recite it perfectly, wait a set interval, then relearn it — and record how much time and effort he saved compared with the first time.

So when his data shows a value of about 34% one day after learning, that does not mean 'you remember 34% of the material.' It means relearning was about a third easier the second time — the savings, not the recall. In his own description, an hour after learning, about half the original work had to be re-expended to get the list back; after eight hours, about two-thirds of the effort had to be made up; and then, in his words, 'gradually the process became slower so that even for rather long periods the additional loss could be ascertained only with difficulty.' Fast at first, then slow — savings of roughly 44% at an hour, about a third at one day, and around a fifth a month later.

Ebbinghaus was also refreshingly humble about his own numbers, warning that given their 'special, individual, and uncertain character,' no one should rush to declare a universal 'law' from them. The shape is robust; the precise percentages are one careful man memorising nonsense in the 1880s, not a constant of human nature. Anyone quoting an exact 'you forget X%' figure as settled fact is overstating what the experiment showed.

A fair question for any 140-year-old finding is whether it survives modern scrutiny. For the forgetting curve, it does. In 2015, Jaap Murre and Joeri Dros published a study in the journal PLOS ONE that set out to replicate Ebbinghaus directly — one subject relearning nonsense syllables at the same intervals Ebbinghaus used (20 minutes, 1 hour, 9 hours, 1 day, 2 days, 6 days, 31 days), over roughly 70 hours of data collection.

Their conclusion was blunt: 'We believe that we may conclude that our attempt to replicate Ebbinghaus' classic forgetting was successful.' The curve is real. They added one intriguing wrinkle: the decline isn't perfectly smooth, and their data showed what looked like a small upward jump around the 24-hour mark — plausibly the effect of a night's sleep helping consolidate the memory. The headline, though, is that the basic shape Ebbinghaus drew by hand in the 1880s reproduced cleanly in the 2010s.

The point of understanding the curve isn't to feel doomed by it — it's to interrupt it. The curve only applies when you do nothing after learning. Each time you successfully retrieve a memory, the next bout of forgetting is slower and shallower, which is why two habits, both strongly supported by research, are the practical answer.

The first is spaced practice — spreading your study over time instead of cramming it into one block. A large 2006 meta-analysis by Cepeda and colleagues reviewed hundreds of experiments and found that distributing study across sessions produces more durable retention than massing it, with the best gap between sessions getting longer as the time until your test grows. A useful note on history: Ebbinghaus observed this spacing effect, but the structured 'spaced repetition' systems people use today came much later — he documented the phenomenon, he didn't build the modern technique.

The second is active recall — testing yourself rather than rereading. A 2013 review by Dunlosky and colleagues evaluated common study techniques and rated practice testing and distributed practice as the two highest-utility methods, while rating rereading and highlighting as low-utility. Put together, the recipe to flatten the curve is simple to state and harder to do: review on a spaced schedule, and make each review an act of retrieval — quiz yourself, don't just look again.

The forgetting curve points to a clear study routine — review on a spaced schedule, and make each review an act of retrieval rather than rereading. The friction is usually building the materials to do that and keeping the schedule. PocketNote auto-generates flashcards and quizzes from your own uploaded notes, PDFs, slides, and lectures, so the self-testing materials come from your actual course rather than a generic deck.

Because the questions are grounded in your sources, each quiz is a retrieval rep on the exact material you need to remember — and spacing those reviews out over days, instead of cramming them into one session, is what turns a steep forgetting curve into a shallow one.