Loading post…
Monthly notes on comprehension, accessibility, and AI tutoring. No spam, no upsells, unsubscribe anytime.
Cramming has a better reputation than it deserves. Pack five hours of study into the night before and you will probably pass the test, which is exactly why the habit survives. The problem shows up later: a month on, most of that material is gone. Readers do the same thing without an exam in sight. We finish a book in a few intense evenings, feel saturated with its ideas, and assume the saturation will last. The research on how memory responds to timing says it will not, and it says so with unusual consistency.
The finding is called the spacing effect: the same amount of study produces far more durable memory when it is spread across multiple sessions than when it is massed into one. It is one of the oldest and most replicated results in experimental psychology, and it comes with something rare in learning science, a usable formula for when to review.
Hermann Ebbinghaus documented both the forgetting curve and the benefit of spaced study in 1885, using himself as the only subject and thousands of nonsense syllables as the material. Memory for new material decays steeply at first, then more slowly, and relearning after a delay takes less effort than the first pass. His methods sound quaint, but his data have aged remarkably well. In 2015, Jaap Murre and Joeri Dros re-ran the original experiment under modern conditions, with one subject spending 70 hours learning and relearning syllable lists at delays from 20 minutes to 31 days. Their replication in PLOS ONE (opens in new tab) closely reproduced the shape of Ebbinghaus' curve. Forgetting is fast, predictable, and apparently a stable fact about human memory across 130 years of cultural change.
The forgetting curve is the half of the story everyone quotes. The other half is what happens when you interrupt the curve with a review. Each well-timed re-encounter slows the next round of forgetting, which is the mechanical core of every spaced repetition system built since.
Individual studies can mislead, so the more useful evidence is the synthesis. In 2006, Nicholas Cepeda, Harold Pashler, and colleagues published a meta-analysis of distributed practice (opens in new tab) covering 839 assessments of the effect across 317 experiments in 184 articles, nearly the entire published literature on the question. Two findings stand out.
First, spacing essentially always helps. Comparing massed presentations against spaced ones across 271 direct comparisons, only a handful showed no benefit or a reversal. Holding total study time constant, splitting it across separate occasions reliably beats one continuous block.
Second, and less widely known, the size of the gap matters, and the right gap is not fixed. The interval between study sessions that produced the best final recall grew as the retention interval grew, meaning the longer you need to remember something, the longer you should wait before reviewing it. That interaction turns spacing from a slogan into a scheduling problem with an actual answer.
To pin the answer down, Cepeda's group ran one of the largest learning-timing studies ever conducted. More than 1,350 people learned a set of obscure facts, reviewed them after a gap ranging from minutes to 3.5 months, and took a final test up to a year later. The results, published as a "temporal ridgeline" of optimal retention in Psychological Science (opens in new tab), are strikingly concrete. With study time held constant, reviewing at the optimal gap instead of immediately produced a 64 percent increase in final recall.
The optimal gaps themselves followed the pattern the meta-analysis predicted. For a test 7 days out, the best gap between sessions was about 1 day. For a test 35 days out, about 11 days. For 70 days, about 21 days. As a rule of thumb, the ideal first review fell at roughly 10 to 20 percent of the retention interval, drifting toward 5 to 10 percent for year-long retention.
Two practical implications follow. The payoff grows with your time horizon: the optimal gap improved recall by 10 percent for a test one week out, and by 111 percent for a test 70 days out. And the penalty is asymmetric: in Cepeda's data, performance fell only gradually as gaps stretched past the optimum, while reviewing immediately forfeited most of the benefit. When in doubt, space wider.
The spacing effect feels backwards because partial forgetting is doing the work. When you review material immediately, it is still fully accessible, processing feels fluent, and your brain treats the repetition as redundant. When you review after some forgetting has set in, retrieval takes genuine effort, and that effort is what strengthens the memory. Researchers call this study-phase retrieval: a spaced review forces you to reconstruct the earlier encounter rather than merely re-perceive it. A second mechanism, encoding variability, adds that spaced sessions happen in different contexts and moods, giving the memory more distinct retrieval routes.
This is the same logic behind the testing effect, where retrieval practice beats re-reading on every delayed measure. The two effects compound: a spaced review that takes the form of a self-test, rather than a re-read, stacks both benefits on the same minutes.
None of this requires software, just a calendar and a little tolerance for the awkward feeling of reviewing half-forgotten material.
Cramming optimizes for the test you take tomorrow. Spacing optimizes for actually knowing things. For readers, who mostly face no test at all, the second goal is the only one that matters.
