Why You Forget What You Learn (and How Spaced Repetition Fixes It)
Finishing a course feels like knowing the material. A month later most of it is gone. That isn’t a personal failing. It is how memory works, and there is a well-tested way to work with it.
Think of a course you finished a year ago, or a textbook you read from cover to cover. How much could you explain right now, without looking anything up? For most of us the honest answer is “the general idea, and not much else”.
Forgetting is not a sign that you learned badly. It is the default. The useful question is what to do about it, and on that question the research on memory is unusually consistent. Two habits make the difference, and both are cheap: test yourself instead of rereading, and spread those tests out over time.
The forgetting curve
In the 1880s the German psychologist Hermann Ebbinghaus memorized lists of nonsense syllables and measured how much he retained after different delays. The result is known as the forgetting curve: memory drops steeply at first and then levels off. Most of the loss happens in the first hours and days.
His second finding matters more. Each time he relearned a list it took less effort, and the forgetting afterwards was slower. Memories that are revisited become more durable. The curve doesn’t just reset. It flattens.
Why rereading feels good and works badly
When you reread your notes, the material feels familiar, and familiarity is easy to mistake for knowledge. Psychologists call this an illusion of fluency. Recognizing a page is a much weaker skill than producing its contents from nothing, and the second one is what you need in an exam, at work, or when the next lesson builds on this one.
Pulling an answer out of your own memory is called retrieval practice or active recall, and it does more than measure what you know. The act of retrieving strengthens the memory. In a well-known 2006 study by Henry Roediger and Jeffrey Karpicke, students who read a passage and then practised recalling it remembered more of it a week later than students who spent the same time rereading it, even though the rereaders were the more confident group.
What spaced repetition is
Spaced repetition combines the two ideas. You test yourself on something, and then you test yourself again later, with the gaps growing each time you succeed: a day, a few days, a week, a few weeks, a couple of months.
The timing is the point. If you review while the memory is still fresh, the review is easy and does little. If you wait until the memory is completely gone, you are learning from scratch. The sweet spot is just before you would have forgotten, when recalling takes effort but still succeeds. The psychologist Robert Bjork called this a “desirable difficulty”: the struggle is what produces the lasting memory.
The benefit of spreading practice out, known as the spacing effect, is one of the most replicated findings in the study of learning. It has been shown for vocabulary, facts, mathematics, motor skills and medical training.
A schedule you can use for anything
| Review | When | What to do |
|---|---|---|
| 1 | The next day | Answer from memory, then check |
| 2 | 3 days later | The same questions, in a different order |
| 3 | A week later | Add one question that connects this topic to another |
| 4 | 2 to 3 weeks later | Explain the idea out loud without notes |
| 5 | 1 to 2 months later | If it is still solid, double the gap from now on |
Treat these gaps as a starting point rather than a law. If a review feels effortless, lengthen the next gap. If you failed, shorten it and go again. Five to ten minutes is plenty for a session. The method works because it is frequent and brief, not because any one session is long.
Recall is not the same as understanding
Flashcards are excellent for facts: vocabulary, formulas, definitions. For ideas they need help, because you can memorize the sentence “amplitudes can interfere” and still be unable to say what a circuit will output.
Three adjustments make spaced practice work for concepts:
- Ask “what happens if” rather than “what is”. A question you have to reason your way through tests the idea, not the wording.
- Predict before you look. Committing to an answer first makes the correction stick. Research on the hypercorrection effect finds that mistakes made with high confidence are the ones most likely to be fixed for good once you see the right answer.
- Rate your confidence. A lucky guess and a sure answer are different evidence about what you know. Tracking the difference tells you what to review.
How we use this in Simple Algorithm
We built our courses around these findings rather than adding them afterwards. Three things happen as you learn:
- Every question asks for your confidence, from guessing to certain. A confident wrong answer is flagged as a misconception instead of being counted as an ordinary mistake.
- Each topic gets an estimate of how well you understand it, using a method from educational research called Bayesian knowledge tracing. A confident correct answer is strong evidence that you know the idea. A lucky guess is weak evidence.
- Each topic also has its own forgetting curve. When the estimated chance that you still remember a topic drops below about three in four, it appears in your review list, and every successful review after a longer gap makes it fade more slowly.
The result is that you don’t plan reviews yourself. You open the review page, answer a few questions on whatever is fading, and close it again. Both of our courses work this way: quantum computing and artificial intelligence.
Common questions
Does spaced repetition work for math and science, or only for memorizing?
It works for both, with different questions. For facts, ask for the fact. For concepts and methods, ask for a prediction or a short problem. What matters is that you produce the answer yourself and that the practice is spread out.
How long should a review session be?
Short. Five to fifteen minutes a day beats an hour once a week, because the benefit comes from the number of well-timed retrievals, not from the time spent.
Is cramming ever worth it?
Cramming works for a test tomorrow and fails for anything after it. If you will need the knowledge again next month, the same hours spread over several days leave you with far more.
What is the difference between active recall and spaced repetition?
Active recall is how you practise: by retrieving the answer from memory instead of rereading it. Spaced repetition is when you practise: at growing intervals. They are separate ideas that work best together.