memori
The short answer
Cram if the test is tomorrow. Space if it is more than a week away. Between those, space if you can. Massed study produces high recall that decays fast; spaced study produces lower peak recall that lasts. Neither is better in the abstract — the answer is fully determined by the delay between studying and being tested.
If you are reading this at 11pm the night before an exam, skip to the one-day protocol. Nothing in the middle of this page will help you tonight, and the advice there is genuinely different from "cram harder".
Cramming works, and here's the evidence
Massed study reliably produces higher immediate recall than spaced study of the same duration. This is not a concession — it is a consistent finding. Rawson and Kintsch found massed readers outperformed spaced readers on a test taken straight after study, and the ordering reversed two days later.
The reason this needs saying up front is that most people arguing for spaced repetition overstate the case, and every student who has ever crammed successfully knows they are overstating it. Once you have told someone that a thing they have personally done many times doesn't work, they stop listening to the part that's true.
What cramming genuinely gives you:
- High retrievability right now. Everything is fresh, primed, and near the top of the stack.
- Coverage under a hard constraint. If there are 200 facts and 8 hours, massed study is the only option that touches all 200.
- A fluency that feels like mastery, which is real for a short window and is why it is so persuasive.
What it does not give you is anything at delay. And the collapse is faster than people expect — it is not that crammed material fades over a term, it is that most of the loss happens in the first week.
Where the crossover actually is
Pulling together the pattern across the spacing literature, holding total study time constant:
| Test happens… | Better approach | Rough size of the difference |
|---|---|---|
| In an hour | Cramming | Clearly better — nothing has decayed |
| Tomorrow | Cramming, slightly | Small; sleep starts to matter more than method |
| In 2–3 days | Roughly even | This is the crossover region |
| In a week | Spacing | Meaningful — commonly 10–20 percentage points |
| In a month | Spacing | Large — spaced groups often retain about twice as much |
| In a year | Spacing | Not really a comparison; crammed material is largely gone |
Two honest caveats about that table. The exact crossover point varies with material and with how the spacing was arranged, so treat "2–3 days" as a region rather than a threshold. And most of these experiments hold total study time constant, which real life does not — a student who crams typically studies less in total than one who spaces, which makes the real-world gap larger than the experimental one.
Every study here compares massed study against spaced study of the same total duration. The comparison that would match most students' actual choice — 8 hours of cramming versus 4 hours spread over a month — is rare in the literature. The spacing effect is strong enough that the spaced condition would plausibly still win at delay, but that is an inference, not a measured result, and it is worth being clear about which is which.
Why spacing wins at delay
Three mechanisms, and they compound.
Retrieval difficulty builds durability. When you study something twice in ten minutes, the second pass is trivially easy — the information is still in working memory and you retrieve nothing. When the gap is three days, retrieval takes effort, and effortful successful retrieval is what strengthens a memory. This is Bjork's desirable-difficulties principle, and it is the same reason active recall beats rereading. It is also, precisely, why cramming feels productive: you have removed the difficulty that does the work.
Encoding variability. Study on Monday, Wednesday and Saturday and you encode the material in three different contexts — different rooms, moods, mental states, adjacent thoughts. That gives the memory more retrieval routes. Eight hours in one library chair on Sunday gives it one.
Consolidation needs sleep, and sleep needs nights. Memory consolidation happens substantially during sleep. Spread study across four days and you get four consolidation cycles operating on the material. Cram it into one day and you get one, which is a hard ceiling no amount of extra hours can raise.
That third point is the one with the sharpest practical implication: the minimum viable spacing unit is a night, not an hour. Two sessions on the same day are barely spaced. Two sessions on consecutive days are properly spaced. If you can only shift your study by one dimension, shift it across nights.
The illusion that keeps people cramming
Cramming feels more effective than it is, and spacing feels less effective than it is. During massed study, material is fluent and familiar, and people read that fluency as learning. During spaced study you keep half-forgetting things, which feels like failure and is actually the mechanism working.
This is a metacognitive error, and it has been measured directly. Learners asked to predict their own future performance systematically over-predict after massed study and under-predict after spaced study. In several experiments participants preferred the massed condition even after being tested and shown that the spaced condition produced better results.
Which means the feeling is not merely unreliable — it points the wrong way. It is worth internalising as a rule:
- If studying feels smooth and you feel like you know it, be suspicious.
- If studying feels effortful and you keep having to dig, that is usually the state that produces durable memory.
The same illusion is what makes rereading and highlighting so persistently popular despite ranking near the bottom of Dunlosky's utility assessment — they maximise the feeling of fluency and minimise retrieval.
If you have one day: the protocol
You cannot space what you don't have time to space. But there is a large difference between good and bad cramming, and the difference is mostly about testing rather than reading.
- Spend 20 minutes making a map before studying anything. List every topic on the syllabus and mark each one: know it, half know it, don't know it. This is uncomfortable and it is the highest-return 20 minutes available to you, because the main way one-day study fails is spending six hours on the first third of the material.
- Cover everything at low resolution first. One fast pass over all of it before any deep pass over any of it. Partial knowledge of all topics scores better than complete knowledge of half, on essentially every exam format.
- Test, don't reread. Close the book and write what you remember; do past questions; explain a topic aloud to an empty room. Retrieval practice beats rereading by a wide margin even within a single session, which makes it the one technique that is right on both sides of the crossover.
- Break the day into 3–4 blocks with real gaps. Morning, afternoon, evening. Hours are much weaker spacing than days but they are not nothing, and revisiting a topic after a four-hour gap is measurably better than revisiting it after four minutes.
- Reserve the last block for the weakest material — but only material you have already seen once. Encountering something genuinely new in the final hour rarely survives to the exam.
- Sleep normally. See below. This is not a wellbeing suggestion, it is a performance one.
- Do a 20-minute retrieval pass in the morning, covering the whole map at high speed. This is the single highest-value block in the entire plan.
Note what is absent: highlighting, making new notes, and rewatching lectures. All three feel productive and all three are close to worthless under time pressure, because none of them involve retrieval.
If you have one week
A week is enough to get most of the spacing benefit, and it changes the plan substantially.
- Days 1–2: pass over everything and turn it into questions — actual cards, or a written question list. Don't try to memorise yet.
- Days 3–5: two retrieval passes daily, split morning and evening. Anything you get right twice in a row, set aside; anything you fail goes to the front.
- Day 6: full mixed pass, everything interleaved rather than grouped by topic. Mixing topics is harder and produces better transfer to an exam, where questions arrive in unpredictable order.
- Day 7: light retrieval on the weakest 20% only, then stop early. Cramming on the final day undoes some of the week's work by displacing sleep and raising anxiety.
Six days of this comfortably beats one day of eight-hour cramming, at roughly half the total hours. If you have a week, you are not in the cramming case at all.
The combination strong students use
Space through the term, then intensify at the end. The final week of concentrated review is not cramming in the pejorative sense — it is raising already-established memories to peak retrievability on a specific date. That works well. What doesn't work is using the final week to build the memories from scratch.
This is the practical resolution of the whole debate, and it is what people who appear to cram successfully are usually doing without describing it that way. They attended, they engaged, they revisited — and then they did an intensive final pass. The intensive pass gets the credit; the four months of distributed exposure did the work.
If you are running a card system, the version of this is concrete: raise your desired retention for the last three or four weeks so intervals shorten and everything gets seen more often near the exam, then put it back afterwards. That gives you the endgame boost without abandoning the schedule.
The reason students default to cramming is rarely ignorance of the research — it's that spacing requires having built something to review, weeks before there is any pressure to. That setup cost is the actual obstacle. It's also the part Memori is aimed at: cards come out of the explanation you were already reading, in a tap, so the distributed half of the strategy doesn't depend on doing deck-building homework in week two of term.
The all-nighter question
Don't. An all-nighter removes the sleep that consolidates everything you studied that evening, and degrades the attention, working memory and reasoning you need the next morning. Studying until your normal bedtime and sleeping beats studying two extra hours and not sleeping, on every measure that matters.
The trade is worse than it looks because you lose on both sides at once. Sleep deprivation impairs the consolidation of what you just learned, and it separately impairs performance during the exam — sustained attention degrades sharply, and so does the ability to hold multiple things in mind, which is most of what a hard exam question requires.
The honest exception: if you are genuinely at zero on a substantial part of the syllabus and the choice is between knowing nothing about it and knowing something in a degraded state, the calculation can change. That is a rescue operation, not a study strategy, and it is worth being clear-eyed that you are trading tomorrow's function for tonight's coverage.
A better version of the same instinct: go to bed at your usual time and get up 90 minutes early. Morning retrieval on consolidated material is worth more per minute than late-night reading on unconsolidated material.
Frequently asked questions
Is cramming better than spaced repetition?
For a test within about 24 hours, cramming is competitive and sometimes better, because massed study produces high immediate recall. Beyond roughly a week, spacing wins decisively and the gap widens. The choice is determined almost entirely by when the test is.
Does cramming actually work?
Yes, for about 48 hours. Articles claiming it never works are overstating the case. What it doesn't do is produce memory that survives to the following week — which is why students who crammed successfully all term often can't answer the same questions at the final.
How long does crammed information last?
Most of the loss happens in the first week. Comparisons typically find massed and spaced close at an immediate test, massed substantially behind at one week, and spaced retaining roughly twice as much at a month for the same total study time.
Should I pull an all-nighter before an exam?
No. It removes the sleep that consolidates the evening's study while degrading the attention and reasoning you need the next morning. Studying to your normal bedtime and sleeping reliably beats two more hours and no sleep.
What's the best way to cram if I only have one day?
Map the syllabus first, cover everything at low resolution before deepening anything, test yourself rather than reread, break the day into blocks with real gaps between them, and sleep normally. Self-testing beats rereading by a wide margin even within a single day.
Can I combine cramming and spaced repetition?
Yes, and it's what strong students effectively do. Space through the term for durable knowledge, then run a short intensive review in the final days to raise everything to peak retrievability. The final push works because it's topping up material that's already there.
Is cramming bad for you if I'll relearn the material later anyway?
It's inefficient rather than harmful. Relearning something you crammed and forgot is faster than learning it fresh — that's the savings effect Ebbinghaus actually measured, covered in the forgetting curve. But you're paying twice for one outcome, and in cumulative subjects like maths or medicine the second payment comes due at the worst moment, when the next topic assumed you still knew the first.
Does this apply to skills as well as facts?
The spacing effect holds for motor skills and problem-solving procedures too, and the crossover is arguably even more one-sided there — massed practice of a skill produces performance during the session that disappears by the next one. Anyone who has crammed a piece of music the night before a lesson has run this experiment.
Where this comes from
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354–380. The 254-study meta-analysis underlying the delay-dependent pattern described here.
- Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. (2008). Spacing effects in learning: A temporal ridgeline of optimal retention. Psychological Science, 19(11), 1095–1102.
- Rawson, K. A., & Kintsch, W. (2005). Rereading effects depend on time of test. Journal of Educational Psychology, 97(1), 70–80. Massed rereading better immediately; spaced better at two days.
- Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249–255. Includes the prediction data showing learners misjudge which condition helped them.
- Dunlosky, J., et al. (2013). Improving students' learning with effective learning techniques. Psychological Science in the Public Interest, 14(1), 4–58. Rates distributed practice and practice testing as high utility, rereading and highlighting as low.
- Bjork, R. A., & Bjork, E. L. (1992). A new theory of disuse and an old theory of stimulus fluctuation. The retrieval-difficulty account of why spacing helps.
- Walker, M. P., & Stickgold, R. (2006). Sleep, memory, and plasticity. Annual Review of Psychology, 57, 139–166. On sleep-dependent consolidation.
