Why you forget 80% of a lecture within a week

You read the chapter three times. You understood it when you closed the book. Ten days later, sitting the exam, all that is left is a vague impression and two keywords.

This is not a lack of attention. It is how a human memory normally works, and it has been documented since 1885.

The forgetting curve

Hermann Ebbinghaus made himself learn lists of meaningless syllables, then measured what remained over time. The result is brutal: without reactivation, you lose roughly half of new information within a day, and most of the rest within the week.

The decay is not linear. It is very steep at first, then it slows. What survived three days stands a good chance of surviving three weeks.

That shape is what makes everything else possible.

Rereading barely works

Here is the uncomfortable part, because it condemns the most common revision method of all.

Rereading a course produces a feeling of mastery without producing memory. The text looks familiar, so you conclude that you know it. But recognising is not recalling: on the day of the exam, nobody shows you the text.

Highlighting fails the same way, only worse — it gives the impression of work while all you did was decorate.

What works: a little discomfort

The effort of retrieval is what engraves. Searching for an answer without seeing it, sometimes failing, then checking — that moment of difficulty is what consolidates.

It is counter-intuitive: a revision session that feels easy is usually a wasted one. A session that costs you is a session that counts.

Hence the flashcard: a question on one side, the answer hidden on the other. The format forces you to produce the answer before seeing it. You cannot cheat on yourself.

The second lever: spacing

Reviewing ten times in one evening is worth less than reviewing five times spread over three weeks. For the same effort, spacing wins by a wide margin.

And Ebbinghaus's curve tells you when to space: just before you forget. Too early and you are reviewing something you already know, wasting time. Too late and you have forgotten, so you start over.

The optimal interval grows as the memory consolidates. One day, then three, then a week, then a month.

The practical problem

You cannot keep that schedule by hand. With four hundred cards, each at its own stage, you would need a spreadsheet updated every evening — and you would give up within three days.

That is exactly what a spaced repetition program solves. Anki works out, for each card individually, the moment you are about to forget it, and shows it to you that day. You open the app, you review what it offers, you close it. The schedule is its problem.

It is free, it is open source, it has existed for fifteen years, and a large share of medical students worldwide rely on it. There is no serious reason to use anything else.

The one real obstacle

Anki does not make your cards. It schedules them, presents them, counts them — but somebody has to write them.

That is where most people stop. Breaking a chapter into atomic cards takes an hour of thankless work, before you have even started revising. You put it off, you never do it, and you go back to rereading your notes while telling yourself it will be enough.

That is the only problem Buvard solves: you drop the chapter in, you get the deck out, you open it in Anki. The thankless hour disappears, the method stays.