Spaced repetition: how to remember 1000 words and never lose one

You learn a new word.

You understand it. You even use it in a sentence.

Three weeks later, someone says it to you and your brain goes: "I know this word…" But the meaning is gone.

This is one of the most frustrating parts of learning a language. The problem is not that you have a bad memory. The problem is that your brain is doing exactly what it was designed to do: forget information that does not seem important.

Spaced repetition works with this process instead of fighting it. The idea is surprisingly simple: review a word around the moment you are about to forget it. Not every day. Not ten times in one evening. At the right time.

This is called spaced repetition, and it completely changes the way you think about learning vocabulary.

Your brain is supposed to forget

In the late 19th century, German psychologist Hermann Ebbinghaus studied how memory changes over time. His work is often associated with the forgetting curve: after learning something new, our ability to remember it can decline quickly, then the rate of forgetting gradually slows down.

Imagine you learn the French word:

Today, it feels obvious. Tomorrow, you probably still remember it. A few days later, you hesitate. Two weeks later: pour… what?

This is normal. If your brain kept every piece of information equally accessible forever, it would have to treat a Wi-Fi password you used once, a stranger's phone number, and the word you need every day as equally important. Instead, memory is selective. Information that returns again and again becomes more important. Information that disappears from your life becomes harder to retrieve.

So forgetting is not necessarily a failure. It is part of the system. The question is: how do we show the brain that a word is worth keeping?

The worst way to learn 50 words

Let's say you have a vocabulary test tomorrow. You open a list of 50 words. You read them. Then you read them again. And again. After an hour, you know almost everything. It feels great. You close the notebook and think: I learned 50 words today.

But there is a problem. You may have learned how to recognize those words right now. That is not the same as building a memory that survives for months. This is why cramming can feel so effective — the information is still fresh, and your brain does not need to work very hard to retrieve it.

Research on the spacing effect has repeatedly found an advantage for distributing learning across time instead of concentrating repetitions close together. A major review by Nicholas Cepeda and colleagues examined hundreds of experiments and assessments of distributed practice.

Five reviews spread across time are usually more useful for long-term memory than five reviews in one sitting.

Learning is not only about the number of repetitions. Timing matters.

The best moment to review a word

Imagine two situations.

Situation A — You learned pourtant five minutes ago. I ask you what it means. Easy. Your brain barely needs to search.

Situation B — You learned pourtant six days ago. You see it again. For two seconds, nothing happens. Then: wait… however!

That second situation is much more interesting. You had to retrieve the word. The memory was becoming less accessible, but it was still there. This is exactly the kind of moment spaced repetition tries to create — not when the word is completely fresh, not necessarily after it has been forgotten for six months, but somewhere near the point where remembering starts to require effort.

This is why spaced repetition can feel strange. Sometimes a word appears and you think: why am I seeing this now? I almost forgot it. Exactly. That is the point.

What is an SRS?

SRS means Spaced Repetition System — a system that decides when you should see a piece of information again. The simplest version might work like this: you learn a word today, remember it tomorrow, and the system waits longer. You remember it again, and the system waits even longer.

A simplified sequence could look like:

These numbers are only an example. A real schedule does not need to follow one fixed sequence. If you keep remembering the word, the intervals grow. If you forget it, the next review comes sooner. So two words learned on the same day can quickly follow completely different paths.

For example, bonjour — you already hear it everywhere, easy, and the system may push it far into the future. But bouleversant — you forget it three times, so the system brings it back sooner.

This is the important part: you should not spend equal time on every word. Your memory does not treat every word equally. Your learning system should not either.

From boxes to algorithms

The idea of spaced repetition existed before smartphones, AI, and learning apps. One of the simplest methods is the Leitner system. Imagine several boxes of flashcards. New and difficult cards stay in the first box and are reviewed frequently. When you remember a card correctly, it moves to the next box. Cards in later boxes are reviewed less often. Forget a card? It moves back. Simple, and surprisingly clever.

But computers can do something better. They can track the history of every individual word: when did you last review it, did you remember it, how difficult was it, how long was the previous interval, how stable does the memory appear to be?

This led to algorithmic spaced repetition systems. One famous example is SM-2, an algorithm developed for SuperMemo. It calculates review intervals using information about previous repetitions and the apparent difficulty of an item. Systems inspired by this family of algorithms became closely associated with flashcard learning and tools such as Anki.

More recently, algorithms such as FSRS — Free Spaced Repetition Scheduler — have taken a more data-driven approach. Instead of treating every learner's memory as identical, FSRS models concepts such as the difficulty and stability of a memory and can optimize parameters using review history.

The mathematics can become complicated. Fortunately, the learner does not need to calculate anything. The experience should be simple: here are the words worth practising today.

Why reviewing every word every day is a terrible system

Imagine you have saved 1,000 words. Reviewing all of them every day is impossible. At 10 seconds per word, one review would take almost three hours. Tomorrow? Three more hours. Congratulations — you have accidentally created a full-time job called "looking at flashcards."

This is where spaced repetition becomes powerful. Most of your 1,000 words do not need attention today:

• Some are strong — you used them recently and remembered them several times. They can wait.
• Some are weak — you keep confusing them. They should return soon.
• A small group may be approaching the point where your memory starts to fail. Those are the interesting words today.

An SRS is essentially a prioritization system for memory. It tries to answer one question: of everything you have learned, what is the most useful thing to review right now? That is a much better question than: which page of my vocabulary notebook should I reread?

But there is one big problem with flashcards

Spaced repetition is powerful, but it can still be used badly. You can create 5,000 cards like this:

and spend years pressing buttons. You may become very good at answering flashcards. That does not automatically mean you can use the language. Words do not live in isolated translation pairs — they live in sentences, situations, stories, conversations.

The French word encore, for example, can be difficult to reduce to one English word. Depending on context, you may encounter ideas like again, still, or another. A flashcard that says "encore = again" is not completely useless, but it is incomplete.

This is why spaced repetition works best when it is connected to real language exposure. You meet a word in a book, a video, a conversation, a podcast — this is where Comprehensible Input helps you discover language in context. Then spaced repetition helps prevent useful words from quietly disappearing from memory.

Input helps you meet language. Spaced repetition helps you keep it.

Recognition is not the same as recall

Here is another trap. Look at this word: pourtant. Now imagine the translation is written directly underneath: however. You look at it and think: "of course, I knew that." Did you? Maybe. But seeing an answer and recognizing it is easier than producing the answer yourself.

Try again. Close the translation. What does pourtant mean?

That tiny moment of searching is important. This is called retrieval practice. Instead of simply exposing yourself to the answer again, you attempt to retrieve information from memory. This is why good practice should sometimes make you think: I know this… give me a second.

Difficulty is not always evidence that learning is failing. Sometimes the effort of retrieval is part of learning. If every exercise feels instantly obvious, you may simply be reviewing information too early or too passively.

So… can you really remember 1,000 words and never lose one?

No. You will forget words. Sorry about the title.

But this is actually the most important lesson in this article. A good learning system should expect you to forget. You are not a database. Memory changes. A word you used every day in university may disappear after five years. A word you forgot three times may suddenly become permanent after hearing it in an emotional conversation. Your vocabulary is alive.

The goal is not to create a perfect archive in your brain. The goal is to make useful language increasingly easier to retrieve when you need it. Spaced repetition does not promise perfect memory. It gives you a smarter way to decide what deserves your attention next. And when your vocabulary grows from 50 words to 500, 1,000, or 10,000, that decision becomes increasingly important.

A simple way to use spaced repetition for language learning

You do not need to become obsessed with algorithms. A practical system can be simple:

The best moment is when a word you practised suddenly appears in a real conversation, and your brain recognizes it instantly.

The future of vocabulary learning is personal

Traditional vocabulary lists assume something strange: that everyone needs the same words, in the same order, at the same time. But your language is shaped by your life.

• A developer may constantly encounter: deployment, dependency, timeout.
• A parent may need: daycare, stroller, tantrum.
• Someone moving to France may quickly learn: justificatif de domicile — not because it was in lesson 17 of a textbook, but because French bureaucracy decided today was the day.

Your vocabulary is personal. Your memory is personal too. So practice should become personal.

The interesting future of spaced repetition is not simply a better flashcard algorithm. It is a learning system that understands which words you care about, which words you already know, which words you keep forgetting, how your memory changes, and what you are likely to need next.

What to check out next

If this topic caught your interest, here are a few great resources: