Distributed Practice vs. Cramming

Take six hours of studying. Spend them in one Sunday block and you will know the material well by Sunday evening and considerably less well by Thursday. Spend the same six hours as one hour on each of six days and you will end up knowing more, remembering it longer, and — this is the part people find hard to believe — you will not have spent a single extra minute. That is the spacing effect: distribution changes the return on identical time.

The reason is that forgetting is not the enemy of learning so much as its raw material. When you return to something after a gap, the memory has faded a little, so getting it back takes effort, and that effort is what strengthens it. Return five minutes later and the material is still sitting there, so nothing is asked of you and nothing is reinforced. Each gap you allow is an opportunity for a productive struggle you would otherwise skip.

This has an unfortunate consequence for how study time feels. The distributed version is less comfortable at every session — you keep starting slightly rusty, you keep noticing what slipped — while the massed version builds a smooth momentum that ends when you close the book. Comfort during study and retention after it are not merely different here; they are close to opposites.

Hour four of a marathon session feels excellent. Terms come instantly, connections seem obvious, and the whole topic feels finally, permanently understood. Every part of that experience is real, and almost none of it is evidence about next week. What you are feeling is the material sitting in easy reach because you touched it forty minutes ago — a state that will not exist on exam day and cannot be tested for from inside the session.

This is the fluency illusion, and cramming produces the strongest version of it available. That matters more than the retention math, because the illusion is what makes the decision for you. Nobody weighs distributed practice against cramming and picks cramming on the merits; they cram, feel genuinely on top of the material, and conclude that it worked. The feedback arrives weeks later, disguised as a bad day rather than a bad method.

The only reliable defense is to stop trusting in-session confidence and check with a delay. Study something today, then tomorrow — before opening anything — write down what you can produce. That gap between what you felt you knew and what actually came back is the number cramming hides.

Being honest about this matters, because advice that treats cramming as simply useless gets ignored by everyone who has passed something by cramming — which is nearly everyone. Massed study does work, on a specific timescale, for specific purposes. Knowing the shape of what it delivers is what lets you use it as a tool instead of falling into it.

Distribution does not require software, a color-coded calendar, or a schedule you will abandon in ten days. It requires that any given topic gets touched on several separate days rather than one long one, and that the gaps grow as the material gets sturdier. Everything else is optional refinement.

The simplest version that holds up: touch new material the day you meet it, again the next day, again about a week later, and once more a few weeks out. Each pass gets shorter, because the passes are retrieval rather than rereading — five honest minutes recalling a topic a week later does more than thirty minutes rereading it that same evening.

Blocking means doing twenty problems of one type before moving to the next. Interleaving means mixing types within a session so consecutive problems rarely match. Blocking feels dramatically better — accuracy climbs within a block, and the sense of getting good at something arrives fast. Interleaving feels worse, produces more mistakes during practice, and generally leads to stronger performance on a later mixed test.

The reason is that blocked practice quietly removes the hardest part of the problem. When every question in a row uses the same method, you never have to decide which method applies; the block already told you. Real exams do not sort problems by technique, so identification is a substantial share of the difficulty, and blocked practice trains everything except it. Interleaving forces you to recognize the problem before solving it, which is exactly what the exam will ask.

The honest nuance is that blocking is not wrong; it is a stage. When a technique is brand new, running several of the same type in a row is how you get the procedure into your hands, and interleaving too early produces confusion without discrimination. The sequence that works: block until you can execute the method without consulting an example, then mix it into a pool with everything else and stop sorting by type. Mixed problem sets and full-length practice sections do this for you automatically.

Then spacing advice about starting three months out is useless to you, and the useful question becomes how to run a compressed period deliberately rather than in a spiral. The goal shifts from building durable knowledge to maximizing what is retrievable on one specific morning, and that changes the priorities in ways worth being explicit about.

Triage first, and triage on two axes at once: how heavily a topic is weighted on the exam, and how weak you currently are on it. Heavy and weak is where every remaining hour goes. Heavy and solid gets a maintenance pass. Light and weak — the topic you have been avoiding all semester, worth three questions — gets abandoned on purpose, and deciding that consciously is what keeps it from consuming an evening by accident.

Then change what you do with the hours. Under time pressure, rereading is the most tempting and least productive option available: comfortable, fast, and leaving nothing behind. Practice questions do triple duty — they retrieve, they reveal gaps precisely, and they rehearse the format. Distribution still applies at this scale, too: a topic revisited on days one, three, and five ends up in better shape than one drilled for six hours on day two.

The bill for a crammed exam is rarely presented on the day. It shows up in the next course that assumes you retained the last one, in the cumulative final that revisits material you technically passed, in the professional exam that samples four semesters at once, and in the strange experience of rereading your own notes and recognizing nothing. Each of those is a relearning cost, and relearning is cheaper than learning but far from free.

This is the strongest practical argument for distribution, and it is about total hours rather than virtue. Spacing is not the more disciplined option you should pick because it is harder; it is the option that returns more retention for the same time and spares you paying twice. In any program with a cumulative exam at the end, that second payment is not hypothetical.