The forgetting curve

In 1885 the German psychologist Hermann Ebbinghaus tested his own memory with lists of nonsense syllables. He found that memory drops steeply in the first hours and days after learning, then levels off. This shape is known as the forgetting curve.

He also found something more useful: every time you successfully recall something, the curve gets flatter. The memory fades more slowly, and the next review can wait longer.

Why spacing beats cramming

Cramming (reviewing the same material many times in one session), can get you through tomorrow’s quiz. But researchers have found again and again that spreading the same study time over several days leads to much better long-term memory. This is called the spacing effect, and it’s one of the most reliable findings in learning science.

Spacing works because of effort. When a memory has started to fade, pulling it back takes work, and that work strengthens it. Reviewing something you saw five minutes ago feels easy, and teaches you little.

Active recall: the other half

Spaced repetition works best with active recall: answering a question from memory instead of re-reading the answer. That’s why flashcards and spaced repetition go together so well. The front asks; you answer; the back checks you.

How the intervals grow

A typical schedule might look like this:

• Learn a new card today.
• Review it tomorrow.
• If you remember it, review again in about 3 days.
• Then about a week later, then a few weeks later.
• If you forget it, it comes back soon and the gaps start growing again.

The exact numbers matter less than the principle: remembered cards wait longer; forgotten cards come back sooner.

What the research says

The spacing effect is one of the most tested ideas in psychology. A 2006 review by Cepeda and colleagues brought together hundreds of experiments with word lists, facts and other material. Spreading study sessions apart beat cramming them together in almost every comparison.

A later study by the same group asked how long the gaps should be. The answer: the best gap grows with how long you want to remember something. If the test is next week, reviewing again after a day or two works well. If you want to know the words next year, the gaps should stretch to weeks.

In 2013, Dunlosky and colleagues rated ten popular study techniques. Only two earned their highest rating: practice testing (answering questions from memory) and distributed practice (spacing). Rereading and highlighting, the two most common habits, were rated low. Flashcards with a review schedule combine both of the winners.

From paper boxes to algorithms

In the 1970s, the German science writer Sebastian Leitner described a simple way to space reviews on paper. Cards live in a row of boxes. A card you answer correctly moves to the next box, which you review less often. A card you miss goes back to the first box. It works, but you have to move every card by hand and keep track of when each box is due.

Software removed that bookkeeping. SuperMemo, written by Piotr Woźniak in the late 1980s, scheduled every card on its own instead of every box. Its SM-2 algorithm became the basis for many apps, including Anki, which in 2023 also added a newer scheduler called FSRS. Toly follows the same principle: each card gets its own next review, based on your answers.

How to start today

1. Keep sessions short. Ten minutes a day beats an hour on Sunday.
2. Add a few new items a day. Reviews grow as your deck grows, so start small.
3. Be honest. If you hesitated or got it wrong, mark it as forgotten. The schedule depends on it.
4. Make cards specific. One idea per card is easier to recall than a paragraph.
5. Don’t skip the hard ones. The cards that feel difficult are the ones doing the most work.

Common mistakes

• Adding too much at once. Every new card brings future reviews with it. Fifty new cards today means a heavy week ahead.
• Grading yourself kindly. If the answer came slowly or only half right, mark it as missed. A generous grade pushes the card too far into the future.
• Skipping a week, then cramming. Missing a day is fine. Come back and clear what’s due a little at a time rather than all at once.
• Stopping at recognition. Recognizing a word when you see it is easier than recalling it when you need it. Say the answer before you reveal it.

Doing it without the admin

Doing spaced repetition by hand (tracking which card is due when), is tedious. Apps do the timing for you. In Toly, you only answer Again or Got it, and every card is scheduled automatically. Toly also writes the cards for you: type a word or a concept, and the back is ready in a second.

Remember it right before you’d forget.

That’s the whole idea. Start with a handful of cards today, review for a few minutes tomorrow, and let the intervals do the rest.

Sources

• Ebbinghaus, H. (1885). Über das Gedächtnis. Leipzig: Duncker & Humblot.
• Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin. doi:10.1037/0033-2909.132.3.354
• Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. (2008). Spacing effects in learning: A temporal ridgeline of optimal retention. Psychological Science. doi:10.1111/j.1467-9280.2008.02209.x
• Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students’ learning with effective learning techniques. Psychological Science in the Public Interest. doi:10.1177/1529100612453266

How many words do you need?

More than most people expect, but far fewer than a dictionary. Research by Paul Nation suggests that reading everyday English comfortably takes around 8,000 to 9,000 word families, and following spoken English around 6,000 to 7,000. A word family counts decide, decided and decision as one.

Coverage matters more than the total. Studies of reading found that learners understand a text well when they know about 98% of its words, roughly one unknown word in fifty. The most frequent words make up most of what you read and hear, so they pay off first. Learn the common words early, then add the words of your own work, studies and interests.

1. Learn words in a sentence

A word on its own is hard to hold on to. A word in a sentence comes with a meaning, a feeling, and grammar. When you learn aprovechar, learn it as “Hay que aprovechar el buen tiempo”: we have to make the most of the good weather. The example does half the remembering for you.

2. Test yourself instead of re-reading

Re-reading feels productive because the words look familiar. But recognising a word is not the same as recalling it. Cover the meaning and try to say it first. This active recall is harder, and that’s exactly why it works.

3. Space your reviews

Review a new word tomorrow, then in a few days, then in a week. Each review, timed just before you’d forget, makes the next one wait longer. This is spaced repetition, and it’s the single biggest upgrade most learners can make.

4. Keep new words to a handful a day

Adding 50 words on a motivated Sunday creates 50 reviews a day later in the week. Five to ten new words a day, every day, builds a much bigger vocabulary over a few months: without burning out.

5. Say it out loud

Hearing and saying a word gives your memory another path to it. Listen to the pronunciation, repeat it, and use it in a short sentence of your own.

6. Connect words to each other

Learn a synonym and an antonym with each new word: ambiguo: vago, claro. Words that are connected to other words are easier to find when you need them.

7. Collect words from real life

The words you meet in a book, a show, or a menu matter to you, and that makes them easier to remember than words from a random list. Save them as you go.

Why some words don’t stick

Some words slip away no matter how often you see them. Three causes come up again and again.

• Look-alikes learned together. Words that look or sound alike, or belong to one tight set such as colors or opposites, get mixed up when you learn them in the same week. A classic 1993 study by Tinkham found that related words learned as a group took longer than unrelated ones. Space them out, and learn one well before adding its twin.
• No picture in your head. Abstract words like consequence are harder than concrete ones like ladder. Tie them to a moment from your own life, or to an example sentence that creates a scene.
• False friends. A word that looks like one in your language but means something else will keep tricking you. Put the difference right on the card.

What to skip

• Long lists. They feel efficient and fade fast.
• Highlighting. It marks words; it doesn’t teach them.
• Cramming before a trip or test. It works for a day, not a month.

A 10-minute daily routine

1. Review what’s due first. Keeping old words costs less than learning new ones.
2. Answer out loud before you check. Recall the meaning, then try the other direction: from the meaning back to the word.
3. Add three to five new words from something you read or heard today.
4. Use one of them in a sentence about your own life, spoken or written.
5. Stop while it still feels easy. Coming back tomorrow matters more than doing more today.

Ten minutes a day adds up to about sixty hours a year: enough for well over a thousand words that actually stay.

Putting it together

A simple routine: collect words during the day, turn them into cards with an example, and review for ten minutes every day. Toly is built for exactly this: type a word, and the card comes with meaning, example, pronunciation, synonyms, and antonyms, then shows up again right before you’d forget it.

Sources

• Nation, I. S. P. (2006). How large a vocabulary is needed for reading and listening? The Canadian Modern Language Review. doi:10.3138/cmlr.63.1.59
• Hu, M., & Nation, I. S. P. (2000). Unknown vocabulary density and reading comprehension. Reading in a Foreign Language, 13(1).
• Karpicke, J. D., & Roediger, H. L. (2008). The critical importance of retrieval for learning. Science. doi:10.1126/science.1152408
• Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin. doi:10.1037/0033-2909.132.3.354
• Tinkham, T. (1993). The effect of semantic clustering on the learning of second language vocabulary. System.

What makes any flashcard work

Two things do the work, whatever the tool. The first is retrieval: answering from memory. In a well-known 2008 study, Karpicke and Roediger had students learn word pairs in a foreign language. Students who kept testing themselves remembered about 80% a week later; students who kept rereading remembered about a third. The second is spacing: reviewing just before you forget. Any method that gives you both will work. The real differences are how long the cards take to make and whether you keep going.

Paper flashcards

Writing a card by hand is a small act of learning in itself, and paper needs no battery. The problem is timing. A stack of paper cards doesn’t know which ones you’re about to forget, so most people review everything or review whatever is on top. Big stacks quickly become hard to manage.

Best for: a short list for a test this week, or young learners.

Paper with a schedule: the Leitner box

Paper cards don’t have to be unscheduled. The Leitner system uses a few boxes: cards you get right move forward to a box you review less often, and cards you miss go back to the first box. It is spaced repetition by hand. The catch is the upkeep: sorting, carrying and remembering which box is due.

Anki

Anki is a free, open-source app that brought spaced repetition to millions. It schedules every card for you and can be customised in almost any way: templates, fields, add-ons, and scheduling options. Thousands of shared decks exist.

The cost is time. You write both sides of every card, and learning the app takes a while. Many people love the method and give up on the card-making.

Best for: learners who want full control and are happy to invest in setup. See our Toly vs Anki comparison.

Anki keeps improving, too. Since 2023 it can use FSRS, a newer scheduler that fits the timing to your own review history. If you already have years of cards in Anki, that alone is a good reason to stay.

AI flashcards

AI flashcard apps write the back of the card for you. In Toly, you type one word, concept, or question and get a finished card (meaning, example, pronunciation, synonyms) scheduled with spaced repetition right away.

The trade-off is control: you edit cards rather than design them. And because AI can make mistakes, it’s worth checking cards for anything that must be exact.

Best for: learners who want the benefits of spaced repetition without the card-making.

Writing good cards, whatever you use

• One idea per card. “Mitochondria make ATP” is a card; a whole paragraph about cells is not.
• Keep the context. A word with an example sentence is easier to recall than the word alone.
• Make it yours. Cards from your own reading and your own mistakes stick better than a list someone else chose.
• Check what matters. AI gets common words right almost every time, but rare words, medical terms and legal terms deserve a quick check against a trusted source.

The time cost, in numbers

Say you add 15 new words a day. Writing a good card by hand, with a meaning, an example and a note on pronunciation, takes about a minute, so that is a quarter of an hour of making before any reviewing starts. Over a month, that is more than seven hours spent building cards. With an AI flashcard app the same cards take seconds each, and the time goes into the part that actually builds memory: the reviews.

Quick comparison

How to choose

If you’ll only study a few cards once, paper is fine. If you love tuning your tools, Anki rewards the effort. If you want to type what you’re learning and simply remember it, an AI flashcard app like Toly gets you there fastest.

Sources

• Karpicke, J. D., & Roediger, H. L. (2008). The critical importance of retrieval for learning. Science. doi:10.1126/science.1152408
• Leitner, S. (1972). So lernt man lernen. Freiburg: Herder.
• Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students’ learning with effective learning techniques. Psychological Science in the Public Interest. doi:10.1177/1529100612453266