Why you forget what you study: The forgetting curve explained

A lecture can make perfect sense on Tuesday and be surprisingly hard to explain by Friday. The slides still look familiar, but close the notes and the details refuse to come back.

That gap between recognizing material and retrieving it is one reason studying can feel more successful than it is. Memory often becomes less accessible when knowledge is not used. Psychologists describe this pattern with the forgetting curve.

The forgetting curve is not a timer counting down to the moment a lecture disappears. It is a general pattern: after learning something, recall often drops quickly at first and then more slowly over time. The speed and shape of that decline depend on the material, the learner, and the way memory is tested.

Understanding that pattern can help with a practical question: when class is over, what should happen next so the material is still available on exam day?

What is the forgetting curve?

The forgetting curve is a graph of how access to a memory changes as time passes without practice. The downward slope represents information becoming harder to recall.

It does not mean the information has been erased. Sometimes an answer feels completely gone until a familiar term, diagram, or first step brings it back. Recognition also tends to be easier than producing an answer with no clues, which is why notes can look familiar even when a blank-page explanation falls apart.

Study websites often attach exact percentages to the curve, such as losing 70 percent of new information within a day. Those numbers are easy to remember, but they are not a dependable forecast for a college lecture. No single percentage applies across a psychology definition, a chemistry calculation, and a historical argument.

A single successful pass through new material does not guarantee that it will remain easy to retrieve.

Where did the idea come from?

The curve came from experiments Hermann Ebbinghaus carried out on himself. He memorized lists of nonsense syllables, waited for different lengths of time, and then learned the lists again.

Instead of asking only whether he could recall a complete list, Ebbinghaus measured how much effort relearning saved. If a list took fewer repetitions to learn the second time, some memory remained even when he could not reproduce all of it. His results showed a steeper decline over earlier delays and a slower decline over longer ones.

The setup was narrow. One person memorizing artificial syllables is different from a class of students learning meaningful course material.

A 2015 replication followed a similar method. One participant spent about 70 hours learning and relearning nonsense-syllable lists after delays from 20 minutes to 31 days. The study produced a broadly similar pattern. It strengthened the evidence for Ebbinghaus's finding under comparable conditions, but it did not turn that pattern into a universal timetable for forgetting lectures.

Why some material fades faster than other material

Think about the difference between a term heard once near the end of class and a concept used in three assignments. Both may have appeared in the same lecture, but they did not receive the same attention, explanation, or practice.

Recall also depends on what the task asks for. Choosing a familiar answer from four options, explaining a process on a blank page, and applying a formula to a new problem place different demands on memory. A topic can feel available in one format and disappear in another.

The question "What percentage of a lecture will I remember tomorrow?" has no single answer. It changes with what was learned, how well it was understood, what happened after class, and how the knowledge must be used.

The forgetting curve describes a tendency, not a fixed outcome. What happens between the lecture and the exam can change what remains accessible.

What helps a memory last?

Two study choices matter here: returning to the material across separate sessions and trying to retrieve it without looking at the answer.

Spacing means distributing study over time instead of packing the same work into one sitting. A 2006 review of distributed practice combined 839 assessments from 317 experiments and found a general retention advantage for spaced study. It also found that useful spacing depends partly on how long the learner needs the material to last.

A 2008 study tested that relationship with more than 1,300 participants. Each person completed an initial learning session, one later review, and a final test. Longer waits before the final test generally favored longer gaps before the review, up to a point.

This does not prove that every student should follow a fixed day 1, day 3, day 7, and day 14 schedule. The experiment tested one review, not a complete semester plan. A 2025 meta-analysis of classroom studies also found an overall benefit for distributed practice, but the size of that benefit varied and the evidence did not identify one best schedule for every course.

The practical lesson is to return to important material in separate sessions and plan those sessions around the date when the knowledge will be needed.

How soon should you review after class?

Separate two jobs that are easy to confuse. First, make sure the course material is accurate. Then test whether it can be recalled.

Soon after a lecture, check the notes while the recording, slides, or reading are available. Correct a misspelled term, a broken equation, or a missing qualification before studying from it. This is source verification, not proof that the material has entered long-term memory.

At a later study session, close the source and try to produce the answer. Use the result to decide when the topic needs attention again:

• If the answer is missing or mostly incorrect, correct it and return after a shorter gap.
• If the main idea is right but important steps are missing, turn those gaps into focused questions for the next session.
• If the answer is complete and can be applied without help, allow a longer gap before testing it again.
• If the exam is near, keep the next attempt close enough to leave time for correction. A distant cumulative exam allows reviews to be spread farther apart.

These are planning rules, not a formula taken directly from one experiment. They use the exam date and the quality of the last attempt to make a reasonable decision for the course in front of you.

For more help building the calendar, use the complete spaced-repetition guide.

Why rereading can give a false sense of progress

Rereading has a purpose. It can restore context, clear up confusion, and help correct an inaccurate note. The problem comes when familiarity with the page is treated as evidence that the answer will be available without it.

Retrieval practice checks that difference. Close the source, answer a question, explain the concept, draw the process, or solve the problem. Then reopen the source and compare.

In a 2006 experiment with prose passages, repeated study led to better recall after five minutes, while free-recall testing led to better recall after two days and one week. A 2021 classroom meta-analysis found an overall benefit from testing across 222 studies, although the results varied with the test format, feedback, timing, repetitions, and comparison activity.

Neither finding means that rereading is useless or that every quiz works equally well. A productive cycle uses both: consult the source to understand and correct, remove the source to retrieve, then check the answer before an error is repeated. The active-recall guide explains how to build stronger retrieval attempts.

Turn one lecture into several memory checks with Bananote

A lecture becomes easier to revisit when the source, notes, and practice questions stay connected. Bananote can handle the setup, while the student still does the remembering and checks the result.

1. Capture the source. Where recording is permitted, record the lecture or upload supported audio or video. A PDF, pasted text, existing notes, or a YouTube link can also provide the source material.
2. Review the generated note. Compare terminology, names, formulas, dates, and course-specific wording with the original material.
3. Choose one small target. "Explain competitive inhibition" is more useful than "study biochemistry."
4. Test it without the note. Answer a flashcard before revealing it, take a scored quiz without consulting the source, or explain the topic from memory.
5. Check the attempt. Correct missed or uncertain answers against the note and the original source. A quiz score reflects the questions asked, so also pay attention to vague explanations and lucky guesses.
6. Plan the next check. Bring weak material back sooner and give a stronger answer more time before the next attempt.

Folders and search can keep each course's material findable between sessions. Use that material to retrieve and correct, rather than treating a generated note as proof that the topic has been learned.

What this looks like for one biochemistry lecture

Suppose a lecture covers a metabolic pathway, its regulatory enzymes, and the conditions that activate each branch. The midterm is five weeks away.

The generated note is checked against the recording and slides. Two enzyme names are corrected before any flashcards or quizzes are created.

During a later session, the note stays closed while the pathway is drawn from memory. Flashcards test the enzyme names, and a quiz covers the regulation rules. Most of the pathway is right, but two regulatory conditions are confused.

The next step is not to reread the whole lecture. The source comes back out long enough to correct those two conditions, which then become the targets for the next attempt.

As the explanation becomes more complete, the sessions can be distributed across the remaining weeks. The schedule is responding to the exam date and the student's performance, not to a universal percentage on a graph.

---

Frequently asked questions

Is the forgetting curve real?

The broad pattern is well supported: newly learned material often becomes harder to recall quickly at first, followed by a slower decline. The exact shape is not identical for every person, subject, or type of test.

How much do students forget after 24 hours?

There is no universal percentage. A number such as "70 percent after one day" leaves out differences in the material, prior knowledge, attention, practice, and the way recall is measured.

Should every lecture be reviewed the same day?

Research does not establish a universal same-day rule. Verify the source while it is available, then schedule closed-note retrieval according to the exam date, the kind of answer required, and the result of the previous attempt.

Is rereading a waste of time?

No. Rereading can help with understanding and correction. It becomes weak evidence of learning when familiar pages replace attempts to explain, recall, or apply the material without help.

How does Bananote help with forgetting?

Bananote can keep supported course sources, structured notes, flashcards, and scored quizzes in one workflow. That makes repeated practice easier to set up. The student still verifies the material, retrieves the answer, corrects errors, and decides when another session is needed.

---

Forgetting after class does not mean the lecture was wasted or that memory is broken. It means understanding something once and being able to retrieve it later are different achievements. Build the second one by returning to the material, testing it without help, and correcting what does not come back.

Try Bananote with one lecture, turn it into practice questions, and use each attempt to decide what needs another review.

Sources

• Murre and Dros: Replication and analysis of Ebbinghaus' forgetting curve
• Cepeda and colleagues: Distributed practice in verbal recall tasks
• Cepeda and colleagues: Spacing effects in learning
• Roediger and Karpicke: Test-enhanced learning
• Yang and colleagues: Testing boosts classroom learning
• Mawson and Kang: Distributed practice in academic settings