The Forgetting Curve: Why You Forget 70% After a Week

The original 1885 experiment

Hermann Ebbinghaus was a German psychologist who decided to study memory using the only subject he could fully control: himself. Over several years in the 1880s, he memorised long lists of nonsense syllables — three-letter combinations like "WID" and "ZOF" that carried no semantic meaning — and then re-tested himself at varying intervals to see how much he could still recall.

The choice of nonsense syllables was deliberate. Real words carry associations (you know what "dog" means, so it is easier to encode), and Ebbinghaus wanted to isolate raw memorisation from prior knowledge. His finding, published in 1885 in Über das Gedächtnis (translated to English in 1913 as Memory: A Contribution to Experimental Psychology), showed that forgetting followed a clear pattern:

• Most forgetting happens in the first hour or two after learning.
• After about a day, the rate of forgetting slows down dramatically.
• Whatever you still remember after a week tends to stick around — at least for nonsense syllables.

This shape — steep initial drop, then a long flat tail — is the forgetting curve. It is the foundation of modern memory research and the reason spaced repetition systems exist.

What the curve looks like

The curve below shows the rough shape Ebbinghaus measured. Retention is plotted against time, with the x-axis compressed (each gridline doubles in time) so the early drop is visible.

The exact numbers depend on what you are learning. Nonsense syllables forget faster than meaningful prose. Material you understand deeply forgets slower than material you only memorised. But the shape — fast initial drop, slow tail — is consistent across replications.

5 modern critiques of the original

Ebbinghaus did important work, but pop-science articles often present his curve as a universal law of memory. It is not. Researchers over the last century have pointed out five important caveats:

1. The curve was measured on one person (Ebbinghaus himself), with nonsense syllables, under controlled lab conditions. Generalising from "the rate at which one 19th-century German forgot ZOF" to "how all humans forget all material" was always a leap. Modern replications use larger samples and varied material, and find broadly similar shapes — but with different numerical values.

2. Material type matters enormously. Meaningful material decays slower than nonsense. Visual material decays slower than verbal. Skills you can perform (riding a bike, writing code) decay slower than facts you can state. The single curve becomes a family of curves, each shifted up or down depending on what you are remembering.

3. Emotional state at encoding changes everything. Information learned while stressed, anxious, or excited gets consolidated differently from information learned in a neutral state. Studying for an exam under deadline pressure is not the same as the controlled lab conditions Ebbinghaus used.

4. Spacing matters more than reviewing. The classic interpretation of the curve says: review before you forget. But research since the 1980s on what Robert Bjork calls "desirable difficulties" shows the opposite — spacing reviews so that some forgetting happens between them actually strengthens long-term memory. The curve is not just a problem; the dip is part of how memory consolidates.

5. Individual differences are large. Some people forget twice as fast as others on the same material. Age, sleep quality, attention, working memory capacity, and prior knowledge all shift the curve. Telling every student "you will forget 70% in a week" misrepresents what is really a population average around a wide individual spread.

Does the curve still hold today?

In 2015, Jaap Murre and Joeri Dros at the University of Amsterdam replicated Ebbinghaus's original experiment with modern methodology and published the results in PLOS ONE (Murre & Dros 2015). One of the authors served as the subject (echoing Ebbinghaus's original design) and learned nonsense syllables under similar conditions. The result: the curve held up. The shape was almost identical to Ebbinghaus's, and the numerical values were within a reasonable margin.

This is a useful piece of evidence, but it has the same limitation as the original — a single subject under lab conditions. Other studies that test groups of students learning real course material find the curve's basic shape but more variation in the numbers. The takeaway: the curve is real, but it is not a precise law you can use to plan study schedules to the hour.

What is robust across decades of replication is the underlying mechanism: without retrieval practice, memories decay over time, and the decay is faster at first than later. This is enough to design effective study strategies, even without precise numbers.

5 techniques that work against forgetting

Knowing the curve exists is not useful on its own. The point is to choose study strategies that flatten it — that slow the decay so more material is still retrievable on exam day. Five strategies have strong evidence behind them.

1. Spaced repetition. Review material at expanding intervals — 1 day, 3 days, 7 days, 14 days, 30 days. Each successful retrieval makes the next forgetting slower. This is the principle behind apps like Anki, Quizlet's spaced mode, and the SRS system inside flashcard tools. Spaced repetition works because it forces your brain to retrieve information just as it is starting to fade, which strengthens the memory more than re-reading does.

2. Active recall over passive reading. Closing your notes and trying to write down what you just read is dramatically more effective than re-reading the same passage. This is well-documented enough to deserve its own article — see active recall vs re-reading for the underlying research.

3. Interleaving topics. Most students study one topic until they feel they know it, then move to the next. Better: rotate between topics within a session. Spending 30 minutes each on three topics, interleaved, beats 90 minutes on one. The discomfort of switching is exactly the kind of "desirable difficulty" Bjork's research identifies as building durable memory.

4. Sleep within 24 hours of learning. Memory consolidation happens largely during sleep, particularly during slow-wave sleep and REM. Studies (Wagner et al. 2004 in Nature, and many follow-ups) show that students who sleep within 24 hours of learning new material retain significantly more than students who stay awake the same amount of time. The all-nighter before an exam is one of the worst-studied common study habits.

5. Elaborative encoding. When you learn something new, link it to things you already know. Explain it to someone (real or imagined). Generate your own examples. Connect it to a story or to an analogy. The more retrieval paths your brain builds to a piece of information, the harder it becomes to fully forget.

What to do with this

The forgetting curve is not a problem to fight in one battle. It is the background condition of how memory works, and the techniques above are not tricks — they are the natural way to align study with the curve instead of against it.

If you want a study tool built around these principles — past exam questions tagged by topic, spaced repetition baked in, and explanations designed for active recall rather than passive re-reading — that is what StudyPilot's question banks are for. Pair them with the 4-step past papers framework and you have a study system that uses the science instead of fighting it.

References

• Ebbinghaus, H. (1885). Über das Gedächtnis. (Translated 1913 as Memory: A Contribution to Experimental Psychology.) Public domain at https://psychclassics.yorku.ca/Ebbinghaus/index.htm
• Murre, J. M. J., & Dros, J. (2015). Replication and Analysis of Ebbinghaus' Forgetting Curve. PLOS ONE, 10(7), e0120644. https://doi.org/10.1371/journal.pone.0120644
• Wagner, U., Gais, S., Haider, H., Verleger, R., & Born, J. (2004). Sleep inspires insight. Nature, 427(6972), 352–355. https://doi.org/10.1038/nature02223
• Bjork, R. A., & Bjork, E. L. (1992). A new theory of disuse and an old theory of stimulus fluctuation. From Learning Processes to Cognitive Processes, 2, 35–67.