The Ebbinghaus Forgetting Curve : understanding and countering natural forgetting

You just finished an exciting book. You came out of an engaging conference. You studied for two hours for an exam. And yet, a week later, you struggle to remember the essentials. A month later, you may only retain a few fragments. It's frustrating, discouraging — and perfectly normal.

This phenomenon was first described with remarkable scientific rigor in 1885 by a German psychologist, Hermann Ebbinghaus. With himself as the only subject of the experiment, thousands of meaningless syllables learned and retested at precise intervals, and a methodology of such rigor that foreshadowed modern standards of experimental psychology, Ebbinghaus uncovered a fundamental law of human memory: forgetting follows a predictable curve, rapid at first and slowing down over time.

More than a century after its publication, Ebbinghaus's forgetting curve remains one of the most important and applicable discoveries in the sciences of memory. Understanding why we forget — and to what extent we can remedy it — is the key to learning more effectively, retaining more, and training our brain throughout the stages of life.

1. Hermann Ebbinghaus, pioneer of memory

Hermann Ebbinghaus (1850-1909) is one of the most original figures in emerging psychology. A contemporary of Wundt — who founded experimental psychology in the laboratory in Leipzig — he chose to work alone, with himself as the subject, on a problem that no one had yet dared to approach rigorously: human memory and forgetting.

His method is both simple and ingenious. To avoid any bias related to the meaning or interest of the memorized materials, he creates meaningless syllables — consonant-vowel-consonant combinations like "DAX", "ZUP" or "LOR" — that evoke no pre-existing associations. He constructs lists of 13, 16, or 20, learns them until he can recite them correctly twice, and then retests them at precise intervals: 20 minutes later, 1 hour, 9 hours, 24 hours, 2 days, 6 days, 31 days.

His measurement index is particularly ingenious: the savings method. Rather than simply counting what he remembers, he measures how many trials he saves during relearning. If learning a list took 20 trials the first time and only 10 the second, the savings is 50% — indicating that 50% of the memory trace is still present, even if it is not consciously accessible. This method captures memory traces that traditional recall tests miss.

2. The forgetting curve: what the data shows

Ebbinghaus's results are striking. After a first learning session leading to perfect memorization of the syllables, here is what he observes during subsequent tests:

The shape of the curve is characteristic: the drop is very rapid in the first hours — about half of the information is lost in the hour following learning — then the curve gradually flattens. After 24 hours, about 65% is forgotten. After a month, only about 20 to 25% of the initial memory trace remains for non-meaningful material.

Before drawing catastrophic conclusions, two important nuances. First, Ebbinghaus worked with meaningless material — intentionally designed not to rely on existing knowledge networks. For meaningful, structured, and emotionally charged material, retention rates are much higher. Second, the savings method reveals that residual traces persist even when conscious recall is impossible — forgetting is not total destruction, but a degradation of accessibility.

3. Why do we forget? The mechanisms of forgetting

The trace degradation theory

The most intuitive theory of forgetting is the passive degradation of the memory trace: over time, the synaptic connections encoding a memory weaken if they are not reactivated. Like a path in a forest that gradually disappears if it is no longer walked, the neural connections representing a memory gradually fade with non-use.

This theory is intuitively satisfying and partially correct — there are indeed active biological mechanisms of "synaptic pruning" during sleep, which selectively reduce the least used connections. But it is not sufficient to explain all forgetting phenomena.

Interference: memories interfere with each other

The interference theory offers an important complementary explanation: forgetting is often caused not by the passive degradation of a memory, but by competition between memories for accessibility. We distinguish between proactive interference (old memories disrupt the retention of new ones) and retroactive interference (new memories disrupt the recall of old ones).

This theory explains why learning similar information in succession — two foreign languages in the same session, two lists of names in the same hour — produces more forgetting than learning different materials. The brain "confuses" similar traces and memories blur each other.

Retrieval failure

A third important mechanism: sometimes, forgetting is not a destruction of the memory but a retrieval failure. The memory trace still exists — as evidenced by Ebbinghaus's savings method — but it is no longer accessible without the right retrieval cue. The phenomenon of "tip of the tongue" is a daily illustration: you know that you know something, but you cannot access it in the moment.

This mechanism explains why retrieval cues — being in the same context as at the time of learning, hearing an associated word, feeling the same emotion — can bring back memories that seemed lost. Memories are not objects that we store and retrieve — they are patterns of neural activation that reconstitute on demand, with more or less ease depending on the available cues.

4. Factors that accelerate or slow down forgetting

5. Spaced repetition: Ebbinghaus's answer

Ebbinghaus did not just describe forgetting — he identified the most effective way to counter it. His concept of spaced repetition is undoubtedly the most practically useful discovery in all of memory psychology.

The fundamental principle

The central idea is counterintuitive: to memorize durably, it is better to review at increasingly spaced intervals than to concentrate all review in a single long session just before the deadline. This superiority of distributed practice over massed practice (all at once) is one of the most robust phenomena in all cognitive psychology — replicated hundreds of times, in very varied contexts.

Here is a concrete example. You learn 50 foreign language vocabulary words in one session. To retain them long-term, two strategies:

"All at once" strategy: review the 50 words 4 times the same evening (2 hours of total work). Result after 1 week: about 50% retention. After 1 month: less than 30%.

Spaced strategy: review the 50 words once on day J, then on J+1, then on J+3, then on J+7 (same total work time, distributed over 7 days). Result after 1 week: about 75% retention. After 1 month: more than 60%.

Spacing produces more durable learning because it generates desirable difficulty — a productive difficulty. With each spaced review, the brain must exert retrieval effort — reactivating a trace that has had time to weaken slightly. This retrieval effort strengthens the trace more effectively than re-reading a still fresh trace.

Optimal review intervals

Post-Ebbinghaus research has clarified the optimal review intervals. A practical rule often proposed is as follows: after initial learning, review at:

• Day 1 (24 hours later): The first review is the most critical. Without it, about 60-70% of the information may be lost. A short review (15-20 minutes) the next day resets the counter and "starts" the forgetting curve from a higher level.
• Day 3 to Day 5: Second review, as the curve begins to decline. Reinforces consolidation.
• Day 10 to Day 14: Third review. At this stage, the trace is sufficiently consolidated for the interval to be extended without risk of massive forgetting.
• Day 30 (one month): Fourth review. The memory trace is now in robust long-term memory for the majority of people.
• Day 90 then every 6 months: Maintenance reviews for information that one wishes to keep indefinitely in active memory.

6. Strategies to consolidate memory

The test as a consolidation tool

One of the most important discoveries in memory neuroscience is the testing effect: self-testing on learned material is significantly more effective for long-term retention than rereading that material the same number of times. When you try to remember something, you are not just checking what you know — you are actively strengthening the memory trace.

This phenomenon is paradoxical for many learners: self-testing is uncomfortable (especially when one fails) and seems to "show" what one does not know, while rereading is comfortable and gives the illusion of knowing. But it is precisely the effort of retrieval — even if unsuccessful — that best consolidates memory.

Elaborate Encoding: Making Sense of Material

The richer a memory is in connections — associated with other knowledge, mental images, emotions, personal examples — the easier it is to retrieve. This is the theory of levels of processing by Craik and Lockhart (1972): "deep" (semantic, elaborative) processing produces a more durable trace than "shallow" (phonetic, visual) processing.

In practice: instead of passively rereading a lesson, try to explain the content in your own words, find personal examples, make analogies with what you already know, and ask yourself "why?" and "how?" rather than "what?". This active processing is more tiring — and much more effective.

Sleep as a Consolidator

Sleep plays a central role in memory consolidation. During sleep, and particularly during slow-wave sleep (SWS), the hippocampus "replays" the sequences of neural activation from the day and gradually transfers them to the cortex — where they are encoded into long-term memory. This process of hippocampo-cortical consolidation is fundamental: learning a lot just before sleeping can be more effective than the same learning in the middle of the day, as the subsequent sleep immediately initiates consolidation.

7. Forgetting Curve and School Learning

Ebbinghaus's forgetting curve should transform the way we think about organizing school revisions — and yet, dominant practices in many educational systems go precisely against what research recommends.

Cramming: Effective in the Short Term, Disastrous in the Long Term

"Cramming" — concentrating all revisions in the 48 hours before an exam — is a universal practice and universally ineffective for long-term retention. It exploits working memory and short-term memory to achieve acceptable performance on exam day — but the information learned this way disappears very quickly afterward, because it never had time to consolidate.

A high school student who crammed for their biology course for the end-of-year exam will likely achieve a decent result — but three months later, when that course is a prerequisite for a biochemistry course, it will be as if they had never learned it. Spaced repetition distributes effort over a longer period for a far superior result in terms of lasting retention.

Adapting Revisions to the Forgetting Curve

For families and teachers, here’s how to implement Ebbinghaus's principles in organizing revisions:

Principle 1 — The revision the next day is mandatory. After each new lesson, a short revision the next day (15-20 minutes) is the most important of all. It can be simple: close the lesson and try to recall the main points, rewrite the key concepts from memory, explain the lesson aloud.

Principle 2 — Gradually space out revisions. After Day 1, revise on Day 4, then Day 10, then Day 25. Spaced repetition applications like Anki automatically manage these intervals based on retrieval performance — freeing up cognitive attention for learning itself.

Principle 3 — Vary the modalities. The first revision can be an annotated re-reading. The second, an attempt to summarize from memory. The third, an oral or written test. Varying modalities strengthens encoding by activating different neural circuits.

8. Taking it on in adulthood: training, languages, skills

The forgetting curve applies to both adults and children. How many expensive professional training sessions leave very little trace six months later? How many language courses abandoned after a few weeks have produced no lasting learning?

Professional training

According to several studies on the effectiveness of professional training, between 60% and 90% of the content of a one or two-day training is forgotten within 30 days — in the absence of any practice or post-training review. This "learning decay" represents a considerable waste of resources.

Organizations that take this issue seriously now integrate post-training "boosts": short questions sent via email or app in the days and weeks following the training, which force participants to retrieve and apply what they have learned. These 5-10 minute boosts dramatically reduce forgetting without requiring a significant time investment.

Language Learning

Language learning is perhaps the context where the principles of spaced repetition are most directly applicable and verifiable. The language learning applications that have had the most lasting success (Duolingo, Anki for specialized vocabulary, Memrise) all integrate spaced repetition algorithms that present each vocabulary item at the optimal time based on the learner's performance.

The superiority of 15 minutes a day for a month over 7.5 concentrated hours in a single weekend is demonstrable in any study on foreign language vocabulary learning. The secret is not the total amount of work — it's the distribution over time.

9. Normal Forgetting vs Pathological Forgetting

The forgetting described by Ebbinghaus is normal forgetting — the result of the memory management processes of a healthy brain. But it is important to distinguish this normal forgetting from pathological forgetting, especially in the context of aging.

What is Normal

Forgetting the name of an occasional acquaintance, having trouble remembering where one has put their keys, no longer recalling the title of a book read two months ago — these are normal manifestations of the limits of human memory, slightly accentuated with age. Even young and healthy people forget a large proportion of what they learn if they do not review it. This is not a sign of pathology — it is the normal functioning of the Ebbinghaus curve.

What Deserves Attention

The signals that deserve medical consultation are different: forgetting important recent events (what one did that very morning), forgetting close people, no longer being able to perform familiar daily activities, or showing significant and progressive changes in personality and functioning. These signs may indicate pathological cognitive decline — Alzheimer's disease or other dementia — which fundamentally differs from normal forgetting in terms of mechanisms and trajectory.

10. Practical Applications and Tools

In 2026, Ebbinghaus's principles are more accessible than ever thanks to digital tools that automate the management of repetition intervals. But technology does not replace the understanding of principles — it amplifies them.

Spaced Repetition Algorithms

The SM-2 algorithm, developed by Piotr Wozniak in the 1980s and used in Anki (the reference flashcard application), is the most well-known implementation of spaced repetition. It assigns each card a review interval based on how easily the user retrieved the information. The easier the retrieval, the longer the interval becomes. In case of an error, the card returns very quickly.

More recent algorithms (FSRS, integrated into Anki since 2023) have improved predictions by relying on data from millions of users and the latest research in cognitive psychology. These tools allow for personalized and optimized learning that even the best student could not reproduce manually.

Applying Ebbinghaus Without Technology

Digital tools are useful, but not essential. Manual spaced repetition — with index cards organized in boxes at multiple intervals (the Leitner method) — is perfectly effective and requires no technology. The key is the principle: review at the right time, with an effort of active retrieval, at progressively increasing intervals.