Shorter Flashcards, Better Recall

My first flashcards looked like this. Front: “What is spaced repetition?” Back: three lines of definition, copied word for word.

Reviews went great. I got almost every card right. Then a friend asked me how the scheduling actually works, and I froze. I could recite the three lines. I couldn’t explain them.

So I went back through my deck, and the problem was obvious. I had been practicing recognition, not recall.

A card that holds too much holds nothing

Memory has an annoying quirk: if something looks familiar, it hands you an “I know this” signal. Put a paragraph in front of you and you scan it, think yep, seen it, and hit “remembered.” The algorithm logs a correct answer and pushes the card out six months. By then you know it even less.

When a card is too long, you’re only answering a fraction of it. The rest gets carried along by that feeling of familiarity.

Three rules I still use

1. One card, one question

Here’s a card I used to have:

Front: What is the mechanism of spaced repetition, and what are its advantages? Back: Based on the Ebbinghaus forgetting curve, it reinforces memory by reviewing just before forgetting. Advantages: efficient, good long-term retention, scales well.

The problem isn’t the answer. It’s the question. “Mechanism and advantages” is two questions, and getting half of it right still counts as right. So I split it:

Front: When does spaced repetition schedule a review? Back: Just before you would forget.

Front: Why does spaced repetition use less total time than cramming? Back: Every review lands at the edge of forgetting, so you need fewer of them.

Now each card can only give one signal: I know it, or I don’t.

2. The front has to fit in one sentence

If I can’t write a short question, it usually means I haven’t understood the material yet. When that happens I don’t make the card. I go back, reread the passage, write one sentence in my own words, and split from there.

A useful test: read the front aloud to someone who hasn’t read the article. Do they know what you’re asking? If not, the question is doing too much work.

3. The back fits on one line

One line is a hard limit, not a style preference. The longer the back, the easier it is to slide past the vague middle — “yeah, roughly that.” A one-line answer forces you to commit. Either you got it or you didn’t, with no gray area.

In my experience, if the answer won’t fit on one line, the card needs splitting.

How long a card should take to write

Thirty seconds.

If I’m still writing after that, I stop and ask myself: did I not understand this, or does the card hold too much? Those need different fixes, but neither one is “keep writing.”

The limit sounds harsh. It’s actually the most useful part, because it turns card-making from a filing task into a comprehension test. If it’s slow, I didn’t get it.

An example from language learning

When I was studying Spanish, my first card for a new word looked like this:

Front: What does sobremesa mean? Back: The time spent lingering at the table after a meal, talking.

I reviewed it for two weeks and still paused whenever I saw it in a book. So I changed it:

Front: Después de comer, nos quedamos de ______ charlando hasta las cinco. Back: sobremesa

The difference: the first card tested whether I remembered an English gloss. The second tested whether I could use the word in a sentence. I needed the second one.

When this doesn’t work

These rules fall apart on material you need to understand as a whole. Reading a paper, what you want to keep is the argument, not one definition. For that I write a long note first, then pull three to five genuinely load-bearing points out of it as cards. Cards hold facts. Notes hold structure. Mixing them ruins both.

The other trap is over-granularity. At one point I made cards for things like who published what in which year. By week three I dreaded opening the deck. My line now: only things I want to be able to say out loud without thinking deserve a card.

What I do now

While reading, I only highlight. No cards. On the weekend I spend twenty minutes going through the week’s highlights, and only the ones that collapse into a single question become cards. The rest stay in the note.

I highlight a sentence in the extension, right-click, and it becomes a card with the front and back already split. The only real advantage is that I don’t switch apps — and when I don’t have to switch, I actually do it.

Last week I deleted a batch of old cards, more than I expected. It stung a little. But review time dropped by nearly half, and I miss fewer of them.

More cards isn’t better. That sounds obvious. It took me two years to actually do it.