The Forgetting Curve, Re-Run 130 Years Later
In 1880 Hermann Ebbinghaus measured how fast he forgot. In 2015 two Dutch researchers repeated his experiment, 70 hours of nonsense syllables, and got almost the same curve, with a twist.

In this post
The forgetting curve is one of the most famous graphs in psychology: memory drops steeply at first, then levels off. It comes from Hermann Ebbinghaus, a German psychologist who in the 1880s spent months memorising lists of nonsense syllables and testing himself. For more than a century, every textbook has reproduced his curve, yet almost nobody had repeated his experiment exactly. In 2015 Jaap Murre and Joeri Dros at the University of Amsterdam did, with Dros as the only subject, just as Ebbinghaus had been his own.
Ebbinghaus’s method
Ebbinghaus wanted material that carried no meaning, so that what he already knew could not help him. He invented nonsense syllables: short strings like a consonant, a vowel and a consonant that are not words. He arranged them in rows of 13 and read each row aloud at a steady pace until he could recite it without error.
To measure forgetting he used a clever idea called savings. After a delay he learned the same row again, and timed how much faster it went the second time. If a row first took 25 repetitions and later took 20, the saving was 20%. Savings can detect memory that is too faint to recall but still makes relearning easier.
The replication
Dros, then 22, followed the method as closely as possible. He learned 70 lists, each made of eight rows of 13 syllables. The syllables were built to suit Dutch, his native language, and any that sounded too much like real Dutch words were removed. Ten lists were relearned at each of seven delays: 20 minutes, 1 hour, 9 hours, 1 day, 2 days, 6 days and 31 days. Learning and relearning took about 70 hours, spread over roughly two and a half months.
The curves
| Delay before relearning | Ebbinghaus, 1880s | Dros, 2015 |
|---|---|---|
| 20 minutes | 58% | 47% |
| 1 hour | 44% | 37% |
| 9 hours | 36% | 28% |
| 1 day | 34% | 32% |
| 2 days | 28% | 23% |
| 6 days | 25% | 17% |
| 31 days | 21% | 4% |
Allowing for individual differences, the two curves are strikingly similar, and the authors found the same shape in two other historical replications. The steep early drop is real: by the one-hour mark, more than half of the work of learning a row had to be done again.
The largest difference is at 31 days, where Dros’s savings were much lower than Ebbinghaus’s. The authors can only speculate about why. The 31-day lists had to be learned early in the experiment, when learning was quicker, and learning time crept up steadily as the weeks went on, perhaps from fatigue or from old lists interfering with new ones. Correcting for that drift raises the figure to about 14%, still well below the others.
The jump at one day
Look again at the table. Savings fall from 1 hour to 9 hours, but then rise a little at 1 day, in Dros’s data and in other replications. Ebbinghaus had noticed this too, and set it aside. A smooth curve should only go down.
The authors conclude that the forgetting curve is probably not completely smooth, and that it shows a jump upwards starting around the 24-hour point. Researchers have long suspected sleep: a one-day delay includes a night’s sleep, and recent sleep research would predict a boost to memory from it. The authors say that this explanation still needs to be established more firmly for this kind of experiment.
Where in the list
The study also looked at position within each row. The first two syllables and the last three were learned and relearned most easily, and showed almost no forgetting; the middle syllables were hardest and were forgotten fastest. This is the familiar pattern of beginnings and endings being easier to remember, though the authors caution that the savings method can exaggerate it for very easy items.
The limits
One person, one kind of material, and that material is deliberately meaningless. Real lessons are full of meaning, connections and prior knowledge, all of which slow forgetting. The exact numbers here do not describe how fast a class forgets a history lesson. What carries over is the shape: fast loss at first, slower loss later, and a bump after sleep.
What the curve means for a lesson
The steep first part of the curve is the part teachers can do most about. If a lot is lost within hours, the most valuable moment to bring material back is soon after it is learned, followed by later reviews at wider gaps.
Working with the curve, not against it
- Close a lesson with a few quick questions on what was just taught.
- Open the next lesson with questions on the last one, after a night’s sleep.
- Return to the topic again a week or so later, then a month later.
- Remember that savings are real even when recall fails: material that seems forgotten comes back faster the second time.
The final point is the encouraging one. A topic that a class seems to have lost completely is not starting from zero. For how to time the reviews, see spacing: three short sessions beat one long one and the nine-year vocabulary study.
Sources
- Murre, J. M. J. & Dros, J. (2015). Replication and analysis of Ebbinghaus’ forgetting curve. PLoS ONE, 10(7), e0120644.
Try it in Giga Quizzes
Free, in the browser, and nothing for your students to install. These guides show each step on the real screens.
New to Giga Quizzes?
Free live quizzes in the browser. Students join from any phone, with no app.
Make your first quiz →

