Spaced Repetition vs Cramming: Which Works Better?

Spaced repetition is the better default when the knowledge needs to survive beyond the next test. Cramming can still be useful when the deadline is tomorrow, but treat it as a salvage plan, not evidence that one long session creates durable learning. If you have several days, split the same study time across separate attempts and make yourself recall the material each time.

Cramming is not so much a study method as a deadline with opinions.

The practical difference

Cramming, called massed practice in research, compresses study into one block. Spaced repetition, or distributed practice, divides it across sessions with gaps between them.

That changes more than the calendar. A massed session keeps the material warm: the page, wording, and last answer are still familiar. A later session makes you find out whether you can reconstruct the idea after it has left the front of your mind. That failed or incomplete attempt is useful because it tells you what the next review should contain.

The fair comparison is not “study versus do nothing.” It is the same amount of planned time arranged in different ways. If the test starts tomorrow, a focused last-minute session can help. If you have a week or a month, spending all the time on the final night gives up several chances to find and repair weak spots.

Question	Cramming	Spaced repetition
When does study happen?	Mostly in one late block	Across separate sessions
What feels easier?	Immediate familiarity	Later recall often feels harder
When do gaps become visible?	Near the deadline	During several earlier returns
Best fit	An unavoidable short deadline	Exams and knowledge needed later

What the research actually supports

A large 2006 review by Cepeda and colleagues examined 839 assessments from 317 experiments. Its useful conclusion was not a magic sequence of days. The gap between sessions and the desired retention interval work together: when people need to remember for longer, the useful spacing pattern changes.

A 2025 meta-analysis of classroom studies narrowed the question to curriculum-relevant learning. It synthesized 22 reports with 31 effect sizes and found a moderate overall advantage for distributed over massed practice. The results also varied substantially across studies. That is a reason to prefer the general pattern, not turn one average into a promise about your grade.

There is one wrinkle at very short delays. UC San Diego researcher Sean Kang notes that some laboratory comparisons favored cramming on tests given within roughly ten minutes of study, but the advantage reversed after a longer delay. Having the material freshly open in your head can help right now. It says less about whether it will still be there later.

Spacing does not tell you what to do inside a session. Rereading on three dates is distributed rereading. A stronger review makes you produce the answer before looking. The active recall guide covers that part; the calendar only decides when you return.

Choose by deadline, not ideology

Start with the deadline: how much time is left, and when must this knowledge still work?

If you have 24 hours

Do not spend the day designing a beautiful month-long system. You do not have a month. Break the available time into a first learning pass and two or three closed-source checks, with real gaps between them. Begin with the highest-value material, and stop adding new material near the end.

This is still short-horizon study. Call it a better-organized salvage plan, not long-term spacing.

If you have 7 days

Use the first session to learn and structure the topic. Return the next day without the source, check the gaps, then review again later in the week. Keep a short final check, but do not make it carry the whole plan.

If you have 30 days

Spend less of the total time on the first exposure and reserve several returns across the month. Longer gaps make each attempt more diagnostic. They also give you time to adjust after a weak review instead of discovering the problem at 11 p.m. before the test.

Three dated calendars for the same topic

Here is a controlled planning example, not a memory experiment. All three calendars use the same six-section course module on database indexes, the same 120 minutes of total study, and the same review action: close the source, explain the mechanisms or work through one query plan, then reopen the source and correct gaps.

The dates start on July 28, 2026. Only the deadline changes.

Deadline in 24 hours: July 29

Date and time	Minutes	Work
July 28, 11:00	45	Learn the six sections and write six plain-language prompts.
July 28, 16:00	25	Answer all six prompts without the source, then mark gaps.
July 28, 21:00	20	Retry the gaps and work through the query plan.
July 29, 08:00	30	Do one complete closed-source explanation, check it, and stop.

The plan creates a few gaps inside the day. It cannot manufacture the longer delays available below.

Deadline in 7 days: August 4

Date and time	Minutes	Work
July 28, 11:00	40	Learn the module and write six prompts.
July 29, 11:00	20	Answer the prompts from memory and correct gaps.
July 31, 11:00	20	Explain the mechanisms in a new order and work the query plan.
August 3, 11:00	25	Do a complete rehearsal and target weak sections.
August 4, 08:00	15	Run a brief final check without adding new material.

Deadline in 30 days: August 27

Date and time	Minutes	Work
July 28, 11:00	35	Learn the module and write six prompts.
July 30, 11:00	20	Answer the prompts from memory and correct gaps.
August 4, 11:00	20	Explain the mechanisms and work the query plan.
August 13, 11:00	20	Do a mixed review beginning with the previous gaps.
August 25, 11:00	15	Rehearse the full explanation without the source.
August 27, 08:00	10	Run a brief final check.

These dates are practical examples, not universal optimal intervals. A spaced repetition schedule should respond to the retention goal and to what happened in the previous review. Weak recall moves the next attempt closer. Clean recall gives you room to wait.

What to do during each review

A gap by itself does no studying. The return still has to ask something of you. Keep the review small:

1. Close the notes, video, or chapter.
2. Answer a few prompts, explain the idea aloud, or solve one representative problem.
3. Reopen the source and compare.
4. Mark what was missing or unclear.
5. Start the next review with those gaps.

This is the same basic separation used in the guide to reviewing notes effectively: clean the source when needed, then make later sessions about recall rather than another tour through every line.

One bad attempt does not mean restarting the whole course. The point of the review is to make the next session narrower.

Where Spacey fits

Spacey handles the timing layer when you still have enough time to distribute reviews. Add a study at the topic level, choose a review plan, and the reviews appear as todos. You do not need to turn the module, book, or lecture into flashcards.

Spacey does not teach the topic, write the six prompts, or decide whether your explanation is correct. It keeps the dates from living in a spreadsheet you will politely ignore.

If the exam is tomorrow, use the time you have. If the knowledge matters next month, start before tomorrow becomes the only option.