The Ebbinghaus Forgetting Curve Experiment
In 1885 a young German philosopher named Hermann Ebbinghaus published a short book, Über das Gedächtnis ("On Memory"), that did something nobody had managed before. It measured forgetting. Not by asking people how well they thought they remembered, but by running hundreds of carefully timed trials and counting. His only participant was himself. His material was a pile of meaningless syllables he had invented. And his central result, the forgetting curve, is still in every introductory psychology textbook almost a century and a half later.
The curve shows that forgetting is fast at first and then slows down. Most of what is lost goes in the first hour or so; after that, what remains fades much more slowly. That simple shape underlies modern ideas about spaced review and helps explain why cramming the night before an exam works for the exam and fails for almost everything afterward.
This article explains what Ebbinghaus actually did, what his numbers mean (they are often misreported), how well the result has held up and where its limits are.
The problem: how do you measure memory?
When Ebbinghaus began in 1879, experimental psychology was barely a field. Wilhelm Wundt opened his psychology laboratory in Leipzig that same year, and Wundt held that higher mental processes such as memory could not be studied experimentally. Ebbinghaus, who had earned a doctorate in philosophy in Bonn, was inspired instead by Gustav Fechner's Elements of Psychophysics, which had shown that sensations could be measured mathematically. Ebbinghaus wanted to do the same for memory.
He faced two problems. First, real material is uneven. A line of poetry might be easy because it rhymes, or because you already know it, or because it moves you. Those differences would swamp any effect of time. Second, he needed a number, an objective measure of how much was retained, that did not depend on his own impression.
The solution: nonsense syllables
For the first problem, Ebbinghaus invented a new kind of material: nonsense syllables. Each consisted of a consonant, a vowel and a consonant, such as "DAX," "BOK" or "YAT." He built a pool of about 2,300 of them, wrote each on a separate slip and drew them at random to make lists. Because the syllables had no meaning, every list started out roughly equally unfamiliar, and lists could be compared fairly.
He was candid that this was not perfect. Some syllables accidentally resembled real German words or were easier to pronounce, and he noted that their difficulty still varied. But they were far more uniform than prose or verse.
The solution: the method of savings
For the second problem, he devised the method of savings. He would learn a list until he could recite it twice without error, recording the time or number of repetitions this took. After a delay, he would relearn the same list to the same standard and record how long that took. If relearning was faster, some memory must have survived, even if he could not consciously recall a single syllable before starting.
Savings were expressed as a percentage. Suppose a list first took 1,000 seconds to learn and, a day later, 660 seconds to relearn. The saving is 340 seconds, or 34% of the original effort.
This method was ingenious because it detects memory that ordinary recall misses. You may be unable to recite a list but still relearn it faster than a fresh one. In that sense, the savings method was an early measure of what is now called implicit memory.
The procedure
Ebbinghaus ran his tests with considerable discipline. He read lists aloud at a steady rate, timed by the ticking of a watch or a metronome. He studied at set times of day, avoided testing when ill or distracted, and noted conditions that might affect performance.
For the forgetting experiments reported in chapter 7 of his book, he used lists of 13 syllables. He learned groups of lists and relearned them after one of seven delays: about 20 minutes, 1 hour, 9 hours, 1 day, 2 days, 6 days and 31 days. Each delay was tested many times with different lists, and the results were averaged. The forgetting series was carried out in 1879–80 and repeated for confirmation in 1883–84.
Walking through a single trial
It helps to picture what one measurement involved. Ebbinghaus would draw 13 syllable slips at random and copy them in order onto a sheet. He then read the list aloud from top to bottom at a fixed tempo, over and over. Every so often he would test himself by trying to recite it from memory, and the moment he could get through it twice without a mistake, he stopped the clock. That learning time was his baseline.
Then he waited: perhaps 20 minutes, perhaps a full day, perhaps a month. During that time he went about his normal life, which inevitably included learning and reading many other things. When the interval was up, he took out the same list and repeated the whole process, again stopping when he reached two flawless recitations. The difference between the two times, divided by the original time, gave the savings score.
Because a single list would give a noisy number, he did this with many lists for each interval and averaged them. He also typically learned several lists in one sitting and relearned them together. The method was laborious. Each data point in the famous table rests on many hours of tedious reading aloud, which is one reason no one attempted a full, faithful replication for well over a century.
Why one subject was enough, and why it was not
To a modern reader, a single-participant study sounds weak. But Ebbinghaus's design had a strength that large group studies sometimes lack: a very large number of trials per condition, from one consistent learner, under tightly controlled conditions. In that respect it resembles the single-subject designs still used in psychophysics, where a few highly practiced observers provide precise data.
The weakness is generality. Ebbinghaus could show how he forgot nonsense syllables, but not how much people differ, or how forgetting depends on age, interest or expertise. Those questions required group studies, which began in earnest in the decades after his book appeared.
The results
These are the savings Ebbinghaus reported, as reproduced in standard historical datasets and textbooks:
| Delay after learning | Savings when relearning |
|---|---|
| 20 minutes | 58.2% |
| 1 hour | 44.2% |
| 9 hours | 35.8% |
| 1 day | 33.7% |
| 2 days | 27.8% |
| 6 days | 25.4% |
| 31 days | 21.1% |
Source: Ebbinghaus (1885), chapter 7. Values are savings in relearning time, not the percentage of syllables he could recall.
Two features stand out. The drop is steep at first: almost half of the original learning effort had to be repeated after just 20 minutes, and more than half after an hour. Then the curve flattens. Between one day and one month, savings fell only from about 34% to 21%. Forgetting was rapid, then slow.
Ebbinghaus fitted the data with a formula in which savings decline with the logarithm of time: b = 100k / ((log t)^c + k), where t is time in minutes and k and c are constants he estimated (about 1.84 and 1.25). Later researchers have proposed many other equations, including power functions and exponentials, and there is no single agreed formula; a common modern finding is that power-type functions fit forgetting data well. Popular articles that give a simple exponential "R = e^(-t/S)" as Ebbinghaus's own equation are wrong; that formulation came later.
A common misreading
The numbers in the table are often presented as "you forget 56% within an hour" or "you remember 33% after a day." Neither is quite right. The figures are savings in relearning effort for meaningless lists learned by one practiced person. They are not percentages of information recalled, and they do not apply directly to meaningful material such as a lecture or a textbook chapter, which is typically forgotten much more slowly. The shape of the curve generalizes widely; the specific numbers do not.
Other things Ebbinghaus discovered
The forgetting curve is the most famous result in the book, but not the only one. Among his other findings:
- List length. He could repeat about seven syllables after a single reading. Beyond that, the number of repetitions needed rose steeply as lists got longer. This is an early observation of the limited span of immediate memory later studied by George Miller and others; see how working memory limits thinking.
- Meaning helps enormously. In a comparison, he memorized stanzas of Byron's poem Don Juan and found each stanza needed only about nine repetitions, far fewer than nonsense lists of similar length. Meaning, rhythm and rhyme are powerful aids to memory, which is exactly why he had avoided them.
- Spacing beats massing. In one well-known comparison with 12-syllable lists, 68 repetitions crammed into a single day left him needing seven more repetitions the next day for an errorless recital, while just 38 repetitions spread over three days achieved the same result. This is the first recorded demonstration of the spacing effect.
- Overlearning. The more repetitions he spent on a list on the first day, the less effort he needed when relearning it the next day. Extra practice beyond what was needed for first mastery continued to pay off.
- Remote associations. When he made lists from syllables that had been one or more places apart in a previously learned list, he learned them faster than fresh lists, suggesting that learning links each item not only to its neighbor but, more weakly, to items further away.
Replication: Murre and Dros (2015)
Because Ebbinghaus studied only himself, a natural worry is whether his curve was a personal quirk. Many partial replications over the 20th century found similar patterns, but few used his exact method. In 2015, Jaap Murre and Joeri Dros of the University of Amsterdam published a direct replication in PLOS ONE.
One participant, Dros himself, spent about 70 hours learning lists of nonsense syllables and relearning them after 20 minutes, 1 hour, 9 hours, 1 day, 2 days or 31 days, closely following Ebbinghaus's procedure. The resulting curve was similar to the original. The authors concluded that the forgetting curve has indeed been replicated.
The study also found something interesting about its shape. In both the original data and the replication, savings at the one-day mark were higher than a smooth curve fitted through the other points predicted. Murre and Dros suggested this "jump" at 24 hours may reflect the consolidation of memories during sleep, which is consistent with a large body of later research on sleep and memory; see how sleep shapes emotional memory.
A replication with a single participant shares the original's limitation, so it confirms that the curve is reproducible under Ebbinghaus's conditions rather than proving that everyone forgets at exactly the same rate. But it is an unusually clean confirmation of a 130-year-old result.
Why does the curve bend?
Ebbinghaus described the shape of forgetting but did not explain it in biological terms, which was impossible in the 1880s. Later researchers have offered several complementary explanations for why early forgetting is so much faster than late forgetting.
One idea is that memories become more stable with age. In 1897 the psychologist Adolf Jost proposed what is now called Jost's law: if two memories are equally strong at a given moment, the older one will fade more slowly. A fresh memory is fragile; a memory that has survived a while has, in some sense, been strengthened. Modern neuroscience describes this process as consolidation, in which a memory trace initially dependent on the hippocampus is gradually stabilized, partly during sleep. The bump at 24 hours in both Ebbinghaus's and Murre and Dros's data fits this picture.
A second idea is about the mathematics. In 1991, John Wixted and Ebbe Ebbesen published an influential analysis in Psychological Science titled "On the form of forgetting," arguing that forgetting data across many experiments are well described by power functions, where the rate of loss falls steadily as time passes. Their reanalysis found that Ebbinghaus's own savings data also decline as a power function of time. A power function naturally produces Ebbinghaus's pattern of steep early loss and a long, shallow tail. Some researchers argue that the power shape emerges from averaging many items that each decay at different rates, so the debate about the "true" equation continues.
A third explanation emphasizes interference. In the first hours after learning, new experiences compete with the fresh memory. Older memories have already survived that competition. This view gained support from a 1924 study by John Jenkins and Karl Dallenbach, in which people remembered more nonsense syllables after a delay spent asleep than after an equal delay spent awake, an effect originally attributed to reduced interference and now linked also to active consolidation.
Common misconceptions about the forgetting curve
"Ebbinghaus showed we forget 70% of what we learn within a day." Figures like this circulate widely in corporate training materials, but they do not come from Ebbinghaus. His figure after one day was about 34% savings for meaningless syllables. The percentage of meaningful information a person can recall a day later varies enormously with the material and the test.
"The forgetting curve is the same for everyone." Ebbinghaus measured one person. The general shape replicates, but slopes vary with individual differences, prior knowledge, attention, sleep and the type of memory test.
"Ebbinghaus invented spaced repetition apps." He discovered the spacing effect and described the benefits of distributed practice, but scheduled flashcard systems were developed a century later, drawing on his work and on many later studies.
"Forgotten means gone." The savings method itself shows otherwise. Even when Ebbinghaus could not recall a list, relearning it was faster than learning a new one. Some trace of the original learning persisted long after conscious recall failed.
Limitations and criticisms
One subject. Ebbinghaus was a highly practiced, motivated learner, and he knew the hypotheses he was testing. Group studies since have found the same general shape, but rates differ across people and materials.
Artificial material. Nonsense syllables were chosen to remove meaning, yet almost everything people want to remember has meaning. Frederic Bartlett, in his 1932 book Remembering, argued that memory for meaningful material is reconstructive, shaped by expectations and prior knowledge, a view supported by later research on false memory. Ebbinghaus's approach tells us little about that side of memory.
Measure matters. Savings, recognition and recall give different curves. Recognition is usually much easier than free recall, and savings can reveal memory that neither detects.
Why do we forget? Ebbinghaus measured forgetting but did not settle its cause. Later work emphasized interference, where other learning disrupts memory, alongside decay and retrieval failure. For an overview of these theories, see why we forget.
From forgetting curve to spaced repetition
Ebbinghaus's spacing observation was largely neglected in classrooms for decades, but research has repeatedly confirmed it. A 2006 meta-analysis by Nicholas Cepeda and colleagues in Psychological Bulletin, covering hundreds of comparisons from verbal learning studies, found that spaced practice reliably produces better long-term retention than massed practice, and that longer retention intervals call for longer gaps between study sessions.
Spaced-repetition software turned this into a tool. Piotr Woźniak's SuperMemo, developed from the late 1980s, scheduled flashcard reviews at increasing intervals timed to anticipated forgetting; open-source programs such as Anki later adopted a modified version of its algorithm. The core idea is that each successful review makes the memory more durable, so the next review can be delayed longer.
Practical lessons that follow from Ebbinghaus and later research include:
- Review soon, then less often. Because forgetting is fastest at the start, a first review within a day of learning does the most good.
- Spread study out. The same total time divided across several days produces better long-term memory than one long session.
- Test yourself. Retrieval practice, recalling information rather than rereading it, strengthens memories more than passive review.
- Make it meaningful. Ebbinghaus showed how much easier meaningful material is. Connecting new information to what you know, as described in the levels of processing framework, slows forgetting.
Ebbinghaus's legacy
Ebbinghaus went on to teach in Berlin, Breslau and Halle, wrote a successful psychology textbook and developed an early sentence-completion test for schoolchildren. He died in 1909. His greatest contribution was methodological. He showed that a higher mental process could be studied with controlled materials, quantitative measures and repeated trials. That approach became the template for the experimental psychology of learning and memory.
Today the forgetting curve is sometimes treated as a precise law with exact percentages. It is better seen as Ebbinghaus himself presented it: a carefully measured description of how one person forgot one kind of material, whose general shape has turned out to be remarkably widespread. For more on learning, see the science of learning.
Frequently Asked Questions
Q: What is the Ebbinghaus forgetting curve? A: It is a graph showing how memory declines over time after learning. Ebbinghaus found that forgetting is rapid in the first hour and then slows, so that much of what remains after a day is still partly retained after a month.
Q: What did Ebbinghaus's experiment show in numbers? A: Using the savings method, he found that relearning a list saved about 58% of the original effort after 20 minutes, 44% after an hour, about 34% after a day and about 21% after 31 days. These are savings in relearning effort, not percentages of syllables recalled.
Q: Did Ebbinghaus really only test himself? A: Yes. He was the only participant, which was common in early experimental psychology. He ran the forgetting series in 1879–80 and repeated it in 1883–84 before publishing in 1885.
Q: Has the forgetting curve been replicated? A: Yes. Many studies have found the same general shape, and a direct 2015 replication by Murre and Dros, using Ebbinghaus's method, produced a closely similar curve, including a bump at 24 hours that may reflect consolidation during sleep.
Q: Why did Ebbinghaus use nonsense syllables? A: To create material of roughly equal difficulty with no meaning or prior associations, so that differences in memory could be attributed to time and practice rather than to how interesting or familiar the material was.
Q: Does the forgetting curve apply to everything I learn? A: The general shape, fast early loss followed by slower decline, applies widely, but the rate depends on how meaningful the material is, how well it was learned and how memory is tested. Meaningful material is usually forgotten far more slowly than nonsense syllables.
Q: How can I beat the forgetting curve? A: Review new material soon after learning it, then at increasing intervals, test yourself rather than rereading, and connect new information to what you already know. Spacing and retrieval practice are among the best-supported study strategies.
Sources
- Ebbinghaus H (1885/1913). Memory: A Contribution to Experimental Psychology (Ruger & Bussenius trans.) — Classics in the History of Psychology
- Murre JMJ, Dros J (2015). Replication and analysis of Ebbinghaus' forgetting curve. PLOS ONE 10(7):e0120644
- Ebbinghaus forgetting data (HistData R package documentation)
- Cepeda NJ et al. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin 132:354–380
- Intro Psych: Ebbinghaus and the savings method
- Forgetting curve overview and formula
- Wixted JT, Ebbesen EB (1991). On the form of forgetting. Psychological Science 2:409–415
- Ebbinghaus, Memory, chapter 8 (spacing and repetition), Classics in the History of Psychology
- SuperMemo: Ebbinghaus and the forgetting curve
Read next:
Follow new articles
Add SciHub101 to your feed reader to see new science and technology stories.
About the Author
SciHub101 is an independent science and technology publication founded and edited by Buğra Sözeri, a software engineer and app developer based in Türkiye. AI tools assist with drafting and research. Source documentation in the existing archive is uneven, and articles are being reviewed and corrected. The Editorial Team byline identifies editorial responsibility; it does not certify specialist or independent human review. Where we compare products, anything built by our own team is labelled as such — see the editorial guidelines and About page for how corrections and disclosures work.
Related Articles

Why Do We Forget? The Science of Memory Decay and Forgetting
Forgetting is not a failure of memory but an adaptive feature of the mind. Information is encoded in the brain through synaptic strengthening, stored in distributed patterns across neural networks, and retrieved by reconstructing those patterns. Forgetting occurs through two mechanisms: decay (the gradual weakening of memories over time if unused) and interference (new information competing with or overwriting old information). The forgetting curve, first measured by Hermann Ebbinghaus in 1885, shows that memories fade rapidly at first, then stabilize. Motivated forgetting — the intentional suppression of unwanted memories — involves the prefrontal cortex inhibiting memory retrieval circuits. Understanding forgetting reveals that memory loss can be adaptive: it reduces cognitive clutter and allows the mind to prioritize relevant, recent information.

Active Recall and Quiz-Based Learning: The Science Behind the Most Powerful Study Method
Decades of cognitive psychology research reveals that the most effective study method is not reading or highlighting. It is active recall: testing yourself on what you have learned, spacing those tests over time, and using retrieval difficulty as the signal for what to study next. The research, pioneered by Roediger and Karpicke, shows that retrieval practice produces substantially larger long-term learning gains than equivalent study time spent on relearning. This article explains the science and shows how to apply it to any subject.

Pavlov's Classical Conditioning: Learning Theory Explained
How Ivan Pavlov's dogs revolutionized understanding of learned behavior and memory formation

The Loftus Misinformation Effect: How Memory Gets Rewritten
Elizabeth Loftus's 1974 car-crash study showed a single word can alter memory. Here's how she proved it, what it means for eyewitness testimony, and why courts still struggle with the findings.