The Forgetting Curve: Why You Forget Things and How to Fight It

Forgetting is not a malfunction — it is what memory does by default, and understanding why makes it far easier to study around.

Where the Idea Comes From

In the late 19th century, a German psychologist named Hermann Ebbinghaus ran one of the first systematic experiments on human memory. He tested himself on lists of meaningless syllables and tracked how much he could still recall after different gaps of time, from minutes to weeks. His results showed a pattern that has held up, in broad shape, across well over a century of later memory research: forgetting is fastest immediately after learning something new, then the rate of loss gradually slows down and flattens out.

The exact numbers Ebbinghaus reported came from one experiment, with himself as the only subject and meaningless nonsense syllables as the material being memorised — not exam content, not anything meaningful or connected to prior knowledge. Precise percentages you sometimes see quoted online (claims like "you forget exactly 50% in a day") are oversimplified at best and frequently just wrong. What has genuinely held up across later, more rigorous research is the shape of the curve, not Ebbinghaus's specific decimal points, so treat any confident-sounding exact number with some scepticism.

The Shape That Actually Matters

For a student, the useful takeaway isn't a number — it's the pattern. Without any further exposure to material after you first learn it, your ability to recall it drops sharply at first and then more slowly over time. This is why a topic you "finished" two weeks ago and haven't touched since can feel almost unfamiliar again when it resurfaces. Your brain isn't being lazy or unreliable. It is doing, more or less, what memory does by default to information that hasn't been asked for again.

Why the Brain Forgets by Default

It can feel like forgetting is a failure of the system, but there's a reasonable case that it's closer to a feature than a bug. A brain that retained every sensory detail of every day with equal strength forever would be retaining enormous amounts of information it will never need again — what you had for breakfast three Tuesdays ago, the exact wording of a sentence you read once and never used. Memory appears to work on something closer to a use-it-or-reinforce-it basis: connections that get reactivated get strengthened, and ones that don't tend to fade in relative accessibility over time.

This is also why retrieval itself is not neutral. Every time you successfully recall something — not just recognise it, but actually produce it from memory — that act of retrieval appears to reinforce the memory trace, not merely confirm that it already existed. This is part of why testing yourself has a meaningfully different effect on long-term retention than passively re-exposing yourself to the same material. See our guide to active recall for how to put that difference into practice.

Why Re-Reading Doesn't Fight the Curve

Re-reading a chapter feels productive, and it isn't useless — it has a real role in building initial understanding of unfamiliar material. The problem is what it does and doesn't do for retention. When you re-read something you've already seen once, your brain recognises it: the words, the layout, the general shape of the argument feel familiar, and that familiarity gets misread as knowledge. But recognising a sentence when you see it again and being able to produce the idea unprompted, days later, with no sentence in front of you, are different skills. Re-reading mostly trains the first one. Little in a typical re-reading session asks you to actually retrieve the information from memory, which is the part that appears to matter most for slowing the forgetting curve down.

How Spaced Review Flattens the Curve

The practical implication of the forgetting curve is straightforward even if it isn't intuitive: a single study session, however thorough, is not enough on its own to keep information accessible for weeks or months. What changes the shape of the curve is reviewing the material again — specifically, reviewing it by trying to retrieve it, not just by looking at it — before you've forgotten it entirely, but not so soon that the review feels trivially easy.

Each successful retrieval appears to do two things: it reinforces the memory, and it also seems to slow down how quickly you'll forget it again, meaning the next useful review point is further away than the first one was. In practice this means your review intervals can reasonably stretch out over time for material you keep successfully recalling — daily, then every few days, then weekly, then longer — rather than reviewing everything on a fixed daily schedule regardless of how well you already know it. For the actual scheduling mechanics of how to structure this kind of spaced review in practice, see our guide to spaced repetition, which covers the how; this article is mainly concerned with the why.

A Practical Starting Rhythm

If you want a starting point rather than a rigid formula: review new material once within the first day, again about a week later if the first review went reasonably well, and then move to a longer interval — roughly monthly — for anything you still need to keep accessible after that. Treat this as a reasonable default to adjust, not a universal law. Material that's conceptually connected to things you already know tends to need less frequent review; material that's arbitrary, like vocabulary, dates, or formulas with no intuitive logic behind them, tends to need more. The right test of whether an interval is working is simple: can you still produce the answer from memory when the review comes around, with genuine effort, not a glance at the first word of the answer?

Material You Understand vs. Material You Memorised by Rote

The forgetting curve doesn't apply equally to everything you study. Isolated, arbitrary facts — a date with no story attached, a formula you've never derived yourself, a term you can't connect to anything you already know — tend to follow something close to the steep version of the curve, because there's nothing else in memory for them to attach to. Material you genuinely understand, in the sense of being able to explain why it's true and how it connects to things you already know, tends to be noticeably more resistant to forgetting, because retrieving any one connected piece can cue the rest of the network back into reach, not just that one isolated fact.

This is a practical argument for spending real time on understanding before treating something as ready for spaced review, rather than jumping straight to memorising it by repetition. A fact you've memorised by rote without understanding it will usually still need more frequent review than one you've genuinely understood, even if both felt equally "learned" on the day you first studied them. If a piece of information keeps needing review far more often than everything around it, that's sometimes a sign it hasn't actually been understood yet, not just that it's inherently harder to remember.

One Common Misunderstanding

A frequent and unhelpful conclusion people draw from the forgetting curve is that if they've forgotten something they studied, the original study session must have failed. That's usually not the right read. Initial learning and long-term retention are genuinely different stages, governed by somewhat different processes. Forgetting a significant portion of something after a single exposure, with no further review, is the normal, expected outcome for almost everyone — it is not a sign that you have a bad memory or that your first study session was wasted. The first session's job was to build the initial understanding; review's job is to keep it accessible. Treating one session as sufficient for both is the actual mistake, not the forgetting itself.

Putting This Into Practice With Quizzes

A low-effort way to build retrieval-based review into your routine, rather than relying on willpower to re-read notes on schedule, is to use short quizzes a few days after you first learn something. A quiz is, by its nature, a retrieval exercise — it asks you to produce an answer, not just recognise one among a reading list. ThinkZene's free practice quizzes are available without signing up, which makes them a reasonably frictionless way to get that kind of spaced retrieval opportunity: pick a topic you studied a few days ago, attempt a short set of questions on it, and treat a wrong or shaky answer as information about what needs another look, not as a bad result.

Frequently Asked Questions

Does the forgetting curve mean I should never re-read my notes?

No — re-reading still has a real role, particularly early on, for building initial understanding of unfamiliar material. The issue is relying on it as your only method for retention over time. Pairing an initial read with later retrieval-based review, rather than more re-reading, is the part that makes the biggest difference. See our guide to active recall for specific ways to practise this.

How fast do I actually forget things?

It varies a great deal by person, by how well the material was understood (not just memorised) in the first place, by how emotionally significant or connected to prior knowledge it was, and by the material itself — a formula with no intuitive logic tends to fade faster than a story or an argument you understood deeply. Don't anchor on a specific percentage you've seen quoted; use the general shape of the curve as a planning principle rather than a stopwatch.

Is there a way to stop forgetting completely?

Not really, and that's not actually a problem worth solving. You don't need everything you've ever studied to stay permanently accessible — review is only useful for the material you actually intend to keep available, whether that's for an upcoming exam or a skill you use regularly. The goal isn't zero forgetting; it's making sure the right things are still there when you need them.

The forgetting curve isn't a problem to solve once — it's a background fact to keep designing your review schedule around, for as long as you're learning anything at all.