The forgetting curve is a graph of how memory fades after you learn something: a steep drop soon afterward, then a slower decline. The APA Dictionary of Psychology describes it as a depiction of forgetting over time, with retention falling suddenly at first and more gradually later. The practical lesson is that when you review matters as much as how long you study. Below you will find what Ebbinghaus actually did, what later research confirms, where the popular version goes too far, and how to turn it into a review schedule.
TL;DR
- Memory loss is fastest right after learning and slows down as time passes. This pattern has held up in replications and in studies of very different material.
- Ebbinghaus tested himself on nonsense syllables, so the exact numbers on popular graphs should not be treated as a rule for your own lectures or textbooks.
- No single equation fits every data set. A large 2024 survey found several functional forms fit best, depending on the study.
- Spacing your reviews and testing yourself both slow forgetting. Cramming can win a test tomorrow but loses over weeks.
- A simple plan: review soon after class, then at widening gaps, using questions and flashcards instead of rereading.
Table of Contents
- What the forgetting curve actually shows
- What Ebbinghaus did, and how it was replicated
- Does it apply to real lectures and textbooks?
- The forgetting curve formula, and its limits
- How spacing flattens the curve
- A review schedule built on the curve
- Our perspective
- Put the curve to work with AtListen
- FAQ
- Sources
What the forgetting curve actually shows
Picture a line that starts high at the moment you finish learning, plunges quickly, then flattens into a long tail. The vertical axis is how much you retain. The horizontal axis is time since learning. That shape is the whole idea.
The National Academies describes an experiment in which people heard a long list of words either once or three times and were tested immediately or after one to four days. The curves showed a large early drop in performance, followed by smaller further losses as the delay grew. The same source states that the rate of loss is greatest early on and diminishes as time goes by, and that this holds across shorter and longer time scales and across different kinds of rememberers.
Two things follow for a student:
- The first hours and days after a lecture are when the most is lost, so that is when a short review pays off most.
- What is left after the steep drop fades slowly. Memories that survive the first stretch can last, especially if you revisit them.
What Ebbinghaus did, and how it was replicated
Hermann Ebbinghaus ran his experiments on himself, memorizing lists of meaningless syllables and later measuring how much effort it took to relearn them. This "method of savings" counts how much faster relearning is than the first round of learning. That saving is the measure of what you still retain.
Because he was his own only subject, people have long asked whether the result is a fluke. A replication by Murre and Dros (2015) used the same savings method. One subject spent about 70 hours learning lists of nonsense syllables and relearned them after 20 minutes, 1 hour, 9 hours, 1 day, 2 days and 31 days. The results were similar to Ebbinghaus's original data.
Still, a single subject is a single subject. The same paper notes that the generality of the original results with other people remains uncertain. That caution matters when you see a tidy graph claiming that you lose a fixed percentage of a lecture by tomorrow.
Does it apply to real lectures and textbooks?
The core shape, fast early loss then slowing, shows up well beyond syllable lists. The National Academies reports a study of factory and office workers trained in CPR. Three years later, skills and knowledge had deteriorated to about 15 percent, following the classic curve of forgetting from three to 36 months. That is real, practical material, not nonsense syllables.
But the exact numbers do not transfer. Educational explainers point out that nonsense syllables are the least applicable to how people actually learn, and warn against applying raw decay rates to lectures or books. Writers focused on learning make a similar point: the curve should not be read as a general graph for everyone, since Ebbinghaus's sample was himself and his content was specific.
So use the curve as a direction, not a prediction. Meaningful material, material you connect to what you already know, and material you have practiced several times will decay differently from a list of random syllables. Your own rate is something you learn by checking yourself.
What changes how fast you forget
A 2016 study by Yang and colleagues had people learn single words and word pairs once, three times or six times, either massed or spread out, and tested them after 10 minutes, 1 day, 1 week and 1 month. They found that recollection decreased significantly and was the main source of forgetting over time, while the feeling of familiarity stayed relatively stable. Distributed learning especially helped memory for associations at longer delays.
That explains a familiar trap. Weeks later, a term can look familiar on the page while you cannot actually recall its definition or explain how it connects to other ideas. Familiarity is not recall, so self-testing is a better check than rereading.
The forgetting curve formula, and its limits
Many explainers present the curve as a neat formula with a time variable and a "memory strength" variable: the stronger the memory, the slower the decline. As a teaching model that is a useful picture. As a law of nature it is too tidy.
A broad 2024 survey by Radvansky and colleagues covered 256 papers and 916 data sets. It found that exponential-power, logarithmic and linear functions best captured the widest range of data, and that a substantial amount of data showed either stable memory over time or even improvement. The authors found it surprising that the power function was not the best-fitting function most often. In plain terms, there is no single equation that every memory follows.
There is a second subtlety. Kornell and Bjork (2025) argue that it is tempting to assume empirical forgetting curves equal the rate at which individual memories are forgotten, but that this assumption would be erroneous. A curve of percent correct depends on how memory strengths are distributed relative to a recall threshold, so the same underlying forgetting can produce differently shaped curves.
What to take from the math:
- Treat any specific percentage on an infographic as an illustration.
- Expect your retention to depend on the material, your prior knowledge and how you studied.
- Rely on the robust finding, early steep loss and slower later loss, to plan your reviews.
How spacing flattens the curve
Ebbinghaus himself concluded that spreading repetitions over time is decidedly more advantageous than massing them at once. Modern reviews agree. Weinstein and colleagues (2018) describe the effect simply: the same amount of studying, spaced out over time, leads to greater long-term retention than the same studying packed into one session. They also report that the benefits of distributed practice were clear in all studies with retention intervals longer than a month.
The proposed reason is intuitive. Spacing lets retrieval strength fade before you restudy, so the effort of bringing the idea back builds storage strength. A review of research in Nature Reviews Psychology names spacing and retrieval practice as two key strategies with research support, and notes that learners often underuse them.
One honest wrinkle: massed practice can produce better recall on an immediate test than spaced practice. That is why cramming feels effective the night before an exam. The payoff of spacing shows up later. If you want to see what that trade looks like over a month, our 30-day cram versus spaced study plan lays it out step by step.

Testing yourself slows forgetting too
Roediger and McDermott (2018) report an experiment where repeated testing during learning led to about 21% greater recall after a week than repeated studying. Their conclusion: retrieval practice enhances long-term retention by slowing the rate of forgetting. Spacing and retrieval work best together: pull the answer out of memory, at gaps that let it fade a little first.
If you want practical formats for this, see our guide to active recall sessions, including short five to ten minute bursts, and our piece on retrieval practice strategies.
A review schedule built on the curve
The National Academies notes that facts crammed at exam time soon vanish unless they are grounded by other study and later reviewed. It also describes distributed practice, reviewing across several days instead of massing trials in one session, producing much better retention after eight days in one experiment.
There is no research-proven magic interval for every student, so use an expanding pattern as a starting point and adjust to how much you actually recall:
- Same day: within a few hours of the lecture, spend 10 minutes turning your notes into questions. Do not just reread.
- Next day: answer those questions from memory. Mark the ones you missed.
- A few days later: test only the missed items plus a handful of others.
- About a week later: a short mixed quiz across topics.
- Before the exam: one or two spaced rounds, longer gaps for what you know well, shorter for what you keep missing.
The point is to space reviews so each one happens when the material has started to fade but is not gone. If you answer instantly, lengthen the gap. If you cannot answer, shorten it.
For lectures and textbooks
Convert each lecture into a small set of questions and definitions right away. Mixing topics in later sessions also helps; our interleaving practice guide covers a 45 minute session. For a wider toolkit, the memory retention techniques for a 7 day push cover the basics, and a memory palace suits ordered lists you must hold in sequence.
For foreign-language vocabulary
Vocabulary is the classic forgetting-curve problem: hundreds of small items, each fading on its own clock. Short daily card sessions beat a long weekly one. Keep each card to one idea, add a sentence in context, and let the cards you know drift to longer gaps. Our flashcard templates for language learners give you formats to copy, and the guide to effective flashcards explains how to write cards that test recall instead of recognition.
Making the schedule stick
A plan fails when you skip it, not when the interval is slightly off. Some students find that studying alongside someone else keeps sessions happening, a habit explored in this piece on body doubling and focus. Pair a fixed daily slot with a short, specific task such as ten questions, and the review habit becomes easier to keep.
Our perspective
We built AtListen around one observation from this research: the hard part of studying is not the first pass, it is coming back to material at the right time, in a form that makes you recall it.
Most students already have the content: lecture recordings, PDFs, textbook chapters, YouTube videos. What they lack is the time to turn it into questions, definitions and flashcards before it fades. So we designed AtListen to take those inputs and produce summaries, key concepts, practice questions with explanations and flashcards built for spaced repetition, in minutes.
We also think depth should be your call. A quick skim before a quiz and a deep pass before a final are different jobs, so depth control lets you set how detailed the study materials should be. Instant feedback on your answers helps you see what you actually know, which is the point of retrieval practice.
The forgetting curve is not a reason to worry. It is a reason to study in smaller, repeated, active sessions, and that is what our tool is designed to make easy.
William
Put the curve to work with AtListen
If you would rather not spend your evenings writing questions by hand, AtListen turns lectures, audio recordings, PDFs, notes, web articles and YouTube links into summaries, key concepts, practice questions and spaced-repetition flashcards. You can start on the Free plan at $0 per month. Pro is $19.99 per month and Max is $39.99 per month, with higher tiers designed for greater usage limits and enhanced features. You can compare them on the AtListen plans page.

For a practical walk-through of building cards, read how to build spaced repetition flashcards with AI, and see how to adjust the depth of your study to match your exam.
FAQ
What is the Ebbinghaus forgetting curve?
It is a graph of memory retention over time after learning. Retention drops sharply soon after learning and then declines more slowly, according to the APA Dictionary of Psychology. Ebbinghaus measured this on himself using nonsense syllables.
Is the forgetting curve still considered accurate?
The basic shape has held up. A 2015 replication produced results similar to the original data. What is less certain is whether the exact numbers apply to other people and to meaningful material, so treat specific percentages as illustrations.
Is there one forgetting curve formula?
No. A survey of 916 data sets found that exponential-power, logarithmic and linear functions fit the widest range of data, and no single simple form fit everything. Use formulas as models, not predictions for your own memory.
How can I slow the forgetting curve?
Space your reviews and test yourself instead of rereading. Research reviews show that spaced study beats the same time spent in one session for long-term retention, and repeated testing produced about 21% greater recall after a week in one experiment. Start with a short review soon after learning, then widen the gaps.
Does cramming ever work?
It can help on a test given right away, since massed practice can beat spaced practice on an immediate test. The cost is that the material fades faster afterward. If you must cram, follow it with short spaced reviews so the knowledge lasts past the exam.
Sources
- APA Dictionary of Psychology: forgetting curve
- Murre and Dros (2015): Replication and Analysis of Ebbinghaus' Forgetting Curve
- National Academies: Memory: The Key to Consciousness
- Simply Psychology: The Forgetting Curve
- Ness Labs: The forgetting curve
- Yang et al. (2016): Effects of learning experience on forgetting rates
- Radvansky et al. (2024): Memory from nonsense syllables to novels
- Kornell and Bjork (2025): Why Empirical Forgetting Curves Deviate from Actual Forgetting
- Weinstein et al. (2018): Teaching the science of learning
- Carpenter, Pan and Butler (2022): The science of effective learning with spacing and retrieval practice
- Roediger and McDermott (2018): Remembering What We Learn
- theNeedle: Body Doubling and focus
