The Forgetting Curve and How Spaced Repetition Beats It

From Tukuk-OS Wiki, the free knowledge base.

Ebbinghaus mapped it in 1885 and the shape has not changed since: memories decay on a schedule you can predict — which means review timed to the decay costs a fraction of the hours cramming takes.

A curve you can set your watch by

In 1885 a German psychologist named Hermann Ebbinghaus memorised thousands of nonsense syllables — WID, ZOF, KEB — and tested himself at increasing delays, inventing the memory experiment single-handedly because he did not trust memories that already had meaning holding them together. What he found became the most reproduced curve in psychology: retention falls steeply at first — a large share of brand-new learning is gone within a day — then flattens. Memory does not leak at a constant rate; it decays fastest when it is youngest.

The curve has one property that turns it from a warning into a plan: every successful review makes the next decay slower. The forgetting that follows the second pass is shallower than the first, the third shallower still. Space your reviews along the flattening curve and each one costs less and buys more.

Why cramming works exactly once, and exactly then stops

Cramming is not a myth — it genuinely moves a test the next morning. Massed practice loads working memory and short-term stores, and if the exam is at 9am you can absolutely ride that to 10. The bill arrives later: the same hours spent in one night are worth a fraction of the same hours spent on tomorrow, next week and next month, because the reviews are what move knowledge into the slow-decay layer. If the subject ends at the exam, cram; if you will need it in the final, in an interview, or in the next course, spacing is not a style preference, it is arithmetic.

The practical intervals that fall out of the research are almost folklore in how consistent they are: review the same day, then the next day, then about a week later, then a month. Each pass can be tiny — a closed-book scribble, a shuffled deck — because the point is the retrieval, not the duration. The companion article on active recall covers why the pull is what strengthens; this one is about when.

Building the schedule without a spreadsheet

Three numbers decide a spaced plan: what day is the exam, how many days are left, and how long one session runs. Everything else follows. If your exam is thirty days out and you give it two hours, the offsets that fit are day 0, 1, 2, 4, 8, 16 and a final recall on the eve — seven sessions, the same total time, and the material is met exactly as it starts to slip. A free study planner does that division for you: enter the exam date and the session length, take the printed dates.

What to do on each pass? The early ones can be first contact — read once, close, write. The middle ones belong to retrieval: run your deck in the flashcards tool and let the knew-it tally sort the live cards from the dead ones. The last one is the dry run: a self-marked quiz under something like exam conditions.

For the thing you cannot reschedule — a reading list, a whole term's revision — the same logic scales up: the weekly study grid interleaves subjects across days so each one is re-met just as it starts to slide, which is spacing applied to a timetable instead of to a card.

The uncomfortable signature of a system that is working

Spaced practice has a tell: it feels harder and slower than re-reading, and the progress is invisible between sessions. That is not a malfunction — the difficulty at the moment of review is the difficulty of the memory being rebuilt, and the gap between sessions is the decay the next review is there to intercept. When a tool shows you a score you earned closed-book, that number is the honest one. It usually looks worse than how studying felt — and predicts far better how the exam will go.

See also

• Spaced-Repetition Algorithms — How SM-2 and Its Successors Compute the Next Interval
• The Leitner System — Flashcard Boxes and Scheduled Review, Explained
• Typing Speed Is a Study Skill — Here Is How to Fix It in Ten Minutes a Day
• Memory Tricks That Actually Work — Acronyms, Chunking and the Ridiculous Image
• How to Plan a Study Week That Survives Real Life
• How Many Pages a Day? The Arithmetic Behind a Reading List That Gets Finished