Flashcards and spaced repetition for MDCAT

Flashcards have a reputation problem. Students remember making them and forget using them. A deck of 500 beautiful cards can sit unopened for a month and still feel like work.

The fix is not a bigger deck. It is a small daily habit, built on two ideas: active recall and spaced repetition.

Why rereading feels productive

Rereading a note feels smooth. The words are familiar, the page looks known, and the brain reads familiarity as understanding.

Recognition is not recall. In an MCQ, you have to produce an answer without the page in front of you. That is a different task, and it needs a different kind of practice.

Active recall is the practice of retrieving an answer from memory. A flashcard is one way to force that retrieval. A blank page and a pen is another.

Active recall in one example

Suppose the card asks: "Which part of the nephron reabsorbs most of the filtered glucose?"

Reading the note again gives you the answer in front of you. A flashcard hides it. You say "proximal convoluted tubule" before you flip. The effort of retrieval is what strengthens the memory, even when you get it wrong and have to correct it.

That is why a card should ask one thing. A card with five facts on the back turns into a reading exercise.

The effort is the mechanism. When retrieval is easy, the review is doing less work. A card you have seen ten times should not feel like a page you have read ten times.

None of this needs a special app. A folded sheet of paper works. The app adds the schedule, which is the part that is hard to do by hand.

What spaced repetition adds

If you review a fact once, you will forget it. If you review it too often, you waste time on facts you already know.

Spaced repetition schedules each review near the point where you might forget. Easy facts wait longer. Hard facts come back sooner. The schedule is personal to the card and to your answers.

This is why a fixed timetable of "revise everything on Sunday" is inefficient. It treats every fact as equally hard and every student as equally ready.

How FSRS schedules a card

mymdcat uses FSRS, a scheduling model that estimates two things for each card: stability and difficulty.

Stability is roughly how long the memory will last. Difficulty is how hard the item is for you. Each time you review, you rate the card. FSRS uses that rating, the time since the last review, and the card's history to update the estimate and choose the next date.

The ratings are:

• Again: you could not recall it.
• Hard: you recalled it slowly or with a hint.
• Good: you recalled it correctly with normal effort.
• Easy: it came back immediately.

The interval is an output of the model, not a rule. FSRS does not put every card on the same review timetable. Two students can rate the same card "Good" on the same day and get different next dates, because their histories differ.

A new card starts with short learning steps. If you rate it Again later, it returns to a short relearning step before it earns a longer interval. The model is adjusting the schedule, not punishing you.

The scheduler also caps an interval at one year, so a card you know well still returns inside the exam cycle.

The 90 percent target is a setting, not a promise

mymdcat asks FSRS to aim for about 90 percent retention. In plain terms, the scheduler tries to bring a card back while there is still a good chance you can recall it.

That is a target for the card schedule. It is not a promise about exam-day recall.

Exam performance depends on coverage, sleep, question style, time pressure and luck. A card can be perfectly scheduled and still fail you in a paper because the paper asks it in an unfamiliar way. The target is a tuning knob, not a promise.

A higher target means more reviews and more time. A lower target means fewer reviews and more forgetting. Ninety percent is a sensible default for a student with a full syllabus.

If the queue feels unmanageable, lower the number of new cards before you lower the target. The target controls how often a known card returns. The new-card limit controls how fast the deck grows.

How to rate a card honestly

The rating is the input. A dishonest rating produces a dishonest schedule.

Rate "Again" when you could not answer, even if the answer feels obvious after the flip. Rate "Hard" when you needed a hint or took a long time. Rate "Good" when you answered correctly with normal effort. Rate "Easy" only when it was immediate.

Do not rate "Easy" to clear the queue. The queue gets shorter today and longer next week, because the model will wait longer before showing the card again.

A daily routine that stays small

Fifteen to twenty minutes a day is a workable start, if the deck is focused.

• Clear the due cards first.
• Add no more than 20 new cards a day.
• Suspend a card that keeps failing after several attempts, and relearn the underlying concept.
• Add cards from wrong answers, not from whole paragraphs.
• Stop when the queue is clear. Do not browse the deck for fun and call it review.

The deck should grow from your errors. A card made from a question you got wrong is already connected to a real exam task.

When the queue is large, do not clear it by rating everything Good. Cut the number of new cards, clear the due cards, and accept that the backlog will take a few days. A backlog is a scheduling problem, not a moral failure.

Where cards fit beside questions and papers

Flashcards are for facts, definitions, formulas and distinctions. They are not a substitute for solving questions.

Use the week like this:

• Cards for the recall layer, 15 minutes a day.
• Practice for application, in 30 to 45 minute blocks.
• Past papers for the mixed, timed layer.

A student who only uses cards can recite definitions and still lose marks on a numerical. A student who only does questions can solve familiar items and forget the definition under pressure. The three layers cover different jobs.

A card can send you back to a question. It cannot replace the question.

The deck is large. You do not need to see all of it today. You need to see the cards that are due.

Five mistakes that waste the deck

1. Writing paragraphs on the back. A card should test one fact.
2. Copying the textbook. Make the card from your own confusion, in your own words.
3. Rating everything "Good". The schedule learns nothing.
4. Cramming the queue the night before a test. Spacing works across days, not across one long night.
5. Using cards as the only practice. Recall is not application.

How we checked this

The FSRS behaviour described here matches the scheduler in the live product: ratings of Again, Hard, Good and Easy, a 90 percent target retention, and a maximum interval of one year. The exam context comes from PMDC notices. We make no claim about a percentage of time saved or a promised score.

Open your due cards on flashcards, or read the SAGE guide for the questions that cards cannot answer.