There is a version of studying that feels productive and is not. It is the evening you spend building four hundred cards for one exam. The deck is beautiful. You never open it again.
Size is the quiet killer of flashcard systems. Not difficulty, not discipline, size. A deck you cannot finish stops being a study tool about three days in, when you realise how far behind you are and quietly stop opening it.
This is how to size a deck to the time you actually have, and what to delete first when it is already too big.
Why an oversized deck fails
A deck has a cost you do not see until you are in it: every card you add is a card you have to review again, and again, on a schedule. Adding cards feels free at the start. It is not. A deck is a debt, and the interest is review time.
When the debt outruns your evenings, the system collapses all at once. You skip a day, then two, then the backlog is so large that opening the app feels like a punishment. The deck that was going to save your exam is the reason you stopped studying for it.
The one rule worth remembering
A deck should not be longer than what you can review in one sitting, times seven.
If you can get through sixty cards comfortably in a session, a deck of a few hundred is fine, because you will see the whole thing across a week. If your honest review capacity is forty cards, then a four-hundred-card deck is ten sessions, and it will be stale before you finish the first pass.
The rule is not about discipline. It is about matching the deck to the loop. A deck is meant to be cycled, and a deck that takes three weeks to cycle is not being cycled at all.
Do the arithmetic for your own schedule
Be honest about the number, because this is where people lie to themselves.
- Time how long it takes you to genuinely answer twenty cards, not skim them. Call that number of minutes your pace.
- Decide how many minutes a day you will actually give this. Not the ideal, the real number.
- Divide. That is your cards-per-day.
- Multiply by seven. That is your deck ceiling for a weekly cycle.
If the arithmetic gives you a ceiling of two hundred and you have built six hundred, the deck is not ambitious. It is a plan to quit, and it will work.
What to delete first
When a deck is too big, do not study harder. Delete. Most generated decks contain a third of cards that ask nothing worth answering, and cutting them is minutes of work against hours of reviewing.
- Cards that restate a heading. If the card is the heading and the answer is the heading, there is nothing to retrieve.
- Cards you already know cold. A card you can answer without thinking is not teaching you anything. Retire it.
- Cards that test trivia. If you cannot imagine the fact being useful, it is scenery.
- Duplicates. Three cards that ask the same thing in different words are one card and two distractions.
Cut to the ceiling, then start. You can always add from the source again later, and you will add the right ones because by then you will know what you do not know.
Why a smaller deck teaches more
Memory responds to retrieval and spacing, not to volume. Two hundred cards you cycle five times beat six hundred cards you cycle once, because the second deck never gets spaced. The gain does not come from how much you wrote down. It comes from how many times you pulled it back out.
This is the same argument behind spaced repetition, and it is why the honest advice on deck size is almost always "smaller than you think."
When a big deck is fine
Large decks are not always wrong. They are fine when they are not meant to be finished in a week: a language deck you add to for a year, a professional reference you dip into, a course deck that grows alongside the course.
The distinction is whether the deck is a loop or an archive. A loop should be small enough to cycle. An archive can be as large as you like, as long as you stop treating it as today's to-do list.
Splitting a deck without losing the thread
The fix for an oversized deck is often to split it, not to shrink it. Break it by chapter or by topic and study one sub-deck at a time. You keep all the material while making each cycle finishable, which is exactly what the rule is asking for.
If you built the cards from one document, the range picker does this for you before you even generate, which is a good habit to get into: pick a chapter, finish it, add the next.
The habit that matters more than size
Every study plan is really a question about the loop. Can you finish the cycle this week, and again next week, without dreading it? If yes, the deck is the right size no matter what number is on it. If no, cut until it is.
What to do when you are already behind
If you are reading this with a deck of a thousand due cards, the advice is not to work harder. It is to reset, deliberately and without guilt.
Start by suspending everything for a day. Not deleting, just taking the deck out of the queue so the app stops shouting at you every time you open it. Then pick the ten most important topics and rebuild a small deck from only those. You are not throwing the material away. You are choosing a new entry point, which is usually the whole problem.
Next, work out what "behind" actually meant. It almost always means you added cards faster than you reviewed them, and the fix is a cap: a fixed number of new cards per day that you know you can clear, set low enough to be boring. A cap feels slow for a week and then feels normal for a year.
Then accept that the missed cards are not all equally lost. Some of what you skipped you already know from other sources, or from the act of writing it down in the first place. The deck was never the only place the knowledge lived. You are not starting from zero. You are starting from a smaller, honest queue, and an honest queue is one you will actually clear.
The deck you can finish this week is worth more than the perfect deck you abandoned last month. Resetting is not failure. Continuing to avoid it is.
Frequently asked questions
Is there a hard number for how many cards is too many?
No fixed number, because it depends on your review speed. The practical ceiling is whatever you can review in one sitting, times seven. Beyond that, the deck stops getting cycled.
Should I delete cards I already know?
Yes. A card you can answer without thinking is not teaching you anything, and it borrows review time from the cards you actually struggle with.
What if my course needs more cards than the rule allows?
Split the deck by chapter and study one sub-deck at a time. You keep all the material and make each cycle finishable.
Do pre-made decks have the same problem?
They have a worse version of it. A shared deck is sized for its author, not you, and it is usually larger than you need. Treat it as a source, not a to-do list.
Does a bigger deck always mean more learning?
No. Two hundred cards cycled five times beat six hundred cards cycled once, because the second deck never gets spaced. Volume is not the mechanism.