The encyclopedia · Product & Design · Product decision · 1994–2011
SuperMemo scheduled reviews right before forgetting, making flashcards permanent
Wozniak's 1994 formula spaced repetitions at the longest intervals that avoid forgetting—learning more in less time.
SuperMemo World
the move
In a 1994 paper in Acta Neurobiologiae Experimentalis, Piotr Wozniak and Edward Gorzelanczyk described the algorithm behind SuperMemo: compute the optimum spacing of repetitions to keep about 95% of learned items in memory.
Seven students memorized more than 35,000 Polish-English word pairs over 18 months. Each repetition was self-graded 0–5; successful recalls lengthened the next interval, failures shortened it, and difficult items were automatically separated into their own schedules.
The result was a simple formula—first interval about five days, later intervals multiplied by a factor tied to each item's easiness—that outperformed both massed and uniformly distributed repetition. The software became SuperMemo, and the spacing idea spread as spaced repetition.
why it works
- Effort is spent exactly where forgetting is imminent
- Self-grading adapts the schedule to item difficulty
- The 95% retention target balances review time against memory
- The formula is simple enough to run in any flashcard app
what transfers
Schedule work by when it is about to be lost: spending the next effort on the item closest to forgetting beats repeating everything evenly.
what came after
SuperMemo grew from a DOS program into a family of algorithms (SM-2 through SM-17) and a commercial product still sold by SuperMemo World in Poznań. The 1994 formula became a reference point for modern flashcard software, making spaced repetition a standard learning technique.
references
- Optimization of repetition spacing in the practice of learning (Acta Neurobiol Exp 54:59-69, updated version)
- Optimization of repetition spacing in the practice of learning (PubMed abstract)
spotted an error? The archive wants to know.