Cramming works. That is why people keep doing it, and any advice that opens by denying it gets ignored by the exact person who needs it.
What cramming does not do is leave anything behind. It is a loan against a specific date, and the repayment is that most of the material is gone within weeks.
What the evidence says
The reference point is the 2006 meta-analysis by Cepeda, Pashler and colleagues, which pooled 839 assessments of distributed practice drawn from 317 experiments across 184 articles. Studying the same material in two separated sessions beat studying it in one continuous block, holding total study time equal, and the advantage held whether the final test came a minute later or over a month later.
Equal time. That is the part that makes it interesting. Spacing is not more work. It is the same work, arranged differently, for a substantially better result.
Dunlosky's 2013 review of ten popular study techniques rated distributed practice as one of only two with high utility across subjects, ages and materials. The other was the testing effect. Rereading and highlighting, the two things students actually do, both landed in the low utility tier.
Why cramming feels better than it is
Because during a cram, everything is fluent.
You have just read it, so it comes to mind instantly, and the mind reads that ease as knowing. It is not knowing, it is availability. Come back in nine days with no cues and the availability is gone.
This is the illusion of competence, and it is the reason the worse strategy reliably feels like the better one in the moment. Spaced study feels harder because you have partly forgotten between sessions and have to rebuild. That difficulty is not a defect in the method, it is the method working, which is the argument in desirable difficulties.
How big should the gap be
The useful finding, and the one almost nobody knows, comes from Cepeda's 2008 study. The optimal gap scales with how long you need to remember.
Roughly, the best interval sits at about 10 to 20 percent of the target retention interval.
| You need to remember it for | Space your sessions about |
|---|---|
| One week | 1 day apart |
| One month | 3 to 5 days apart |
| Six months | 3 weeks apart |
| One year | 1 to 2 months apart |
Two implications worth sitting with. If your exam is Friday and it is Monday, daily sessions are close to optimal, and you do not need an elaborate algorithm. And if you want the material for a career rather than a semester, the gaps should be far longer than any student would voluntarily choose, which is precisely why spaced repetition software exists: to schedule intervals your instincts would never pick.
When cramming is the correct decision
Sometimes it is. Be honest about which case you are in.
Cram when the deadline is real and the material is disposable. A compliance quiz, a one-off certification, a subject you will never use again. Optimising for long-term retention of something you do not want to retain is a waste.
Cram when you have already run out of time. Three days out, spacing advice is a lecture about a decision you can no longer make. The question is now how to cram well.
Do not cram material that later material builds on. Anatomy before physiology, calculus before mechanics, the fundamentals of a language. Crammed foundations mean paying twice, and the second payment is larger.
How to cram properly
If you are here anyway, some crams are much better than others.
Self-test, do not reread. Even inside a single session, retrieval beats review. Close the notes, write what you remember, then check. See active recall.
Split the block. Three separated ninety minute sessions beat one four hour block, even if all three happen in one day. You get a real, if reduced, spacing effect from gaps measured in hours.
Sleep between the last session and the test. Consolidation is not optional and it is not negotiable. An all-nighter trades a memory process you cannot replace for revision hours you probably could. See sleep and memory.
Interleave the topics. Mixing problem types within a session outperforms blocking them, because it forces you to identify which method applies rather than executing a method you were just handed. See interleaving.
Write for retrieval next time, not for coverage now. If your notes come out of a cram as transcription, they are useless in three weeks. Questions with hidden answers survive. See how to study from notes.
The honest summary
Cramming and spacing are not competing philosophies. They are tools with different jobs.
Cramming buys performance on a date. Spacing buys retention over a period. The mistake is not using the first one, it is using the first one on material you actually wanted to keep, and then being surprised when it is not there.
More in the spacing effect, the forgetting curve, and how to take notes for exams.