A term of cards for one module, banded and kept together with the ones that keep failing on top.
Cards are the most used revision tool in existence and the most casually made. The usual pack is built by working through a glossary: term on the front, definition on the back, three hundred cards by the end of a fortnight. That pack tests one thing, which is whether you recognise a word, and exams almost never ask that.
The card is a question. Everything that determines whether the pack is worth carrying is decided by how that question is phrased, and the fix costs nothing except the willingness to write a longer front.
Recognition against production
Turn a card over and read the definition, and there is a strong feeling of having known it. That feeling is recognition, and it appears reliably for material you could not have produced unaided thirty seconds earlier. Recognition is a weaker form of memory than production, and it collapses in the exam hall where nothing is offered to recognise.
The test is whether you could have written the back before turning the card. Not sketched, not gestured at: written. If the honest answer is no, the card was scored wrongly, and a pack scored wrongly for two months produces very high confidence and a mediocre paper.
The rewrite is mechanical. Take a front that says "opportunity cost" and change it to "why is the opportunity cost of an hour of revision not zero for a student with no job". Take "mitosis" and change it to "what would go wrong if the spindle checkpoint failed". The topic is identical. The demand is completely different, and the second version cannot be answered by recognition.
A front that takes two lines to state, which is normal once the card asks for an explanation.
Four kinds of front worth writing
Explanation fronts ask why or how, and are the backbone of the pack. Application fronts give a short scenario and ask what happens, which is the closest a card gets to an exam question. Contrast fronts name two things that students routinely confuse and ask for the distinction that matters, which is where a surprising number of marks live. Failure fronts ask what breaks if a condition is removed, and they test whether you understand the mechanism or have memorised its output.
Keep one demand per card. A front asking for three causes will fail whenever you produce two, and a card that always fails teaches you nothing except discouragement. Split it into three cards, or make the front ask for the strongest cause and why it is the strongest.
Cards can carry images, and for some subjects they must. An unlabelled diagram on the front with the labels on the back is a genuine production task. So is a graph with the axes given and the curve missing.
Sorting after a session: the small pile is the cards that failed and will be seen again tomorrow.
Making the pack without losing a week
Cards get made continuously or they get made in a panic. The sustainable version is fifteen minutes after each lecture, writing five to ten cards from that hour's material while the confusions are still fresh; the marked gaps described in rewriting notes is not revision convert into card fronts almost directly.
Do not use a downloaded pack as your main source. Someone else's cards encode someone else's understanding and, more importantly, someone else's confusions. The act of writing the front is where a large share of the benefit sits. A shared pack is useful for checking coverage after your own is built.
Once the pack exists it needs a schedule, and the schedule is the difference between a pack you own and a pack you use. The paper version, with the boxes and the two clause rule, is set out in spaced repetition without an app.
Before a card joins the pack, check it against these:
- The front asks for an explanation, an application, a distinction or a consequence.
- There is exactly one thing to produce on the back.
- You wrote it yourself, from a lecture or a mistake, not from a glossary.
- You answer aloud or on paper before turning it, and score honestly.
Now the limits, because cards are oversold. A pack cannot teach structure. It will not help you decide what an essay question is really asking, how to order an argument, or which of six relevant points deserves two paragraphs and which deserves a clause. Those are composition skills, and they are trained by writing plans and full answers against the clock, as described in reading the mark scheme before you answer.
Cards also carry a real risk of context dependence. Answering "why does elasticity change along a straight demand curve" fluently in your bedroom is not the same as recognising that an exam question about a firm's pricing decision is asking exactly that. Mixing cards from different topics in one session, rather than working through them in neat topic blocks, blunts that problem considerably, which is one of the arguments in interleaving two subjects in one sitting.
Finally, a pack has an upper size. Past roughly two hundred cards per module the daily pass stops fitting into a day, and cards start being skipped silently rather than deliberately. When that happens, retire the cards that have passed three monthly reviews. They are done, and keeping them in circulation costs you the attention that belongs to the twenty cards you keep failing.