How Spaced Repetition Makes Long-Term Memory More Durable

A Nineteenth-Century Psychologist's Discovery Still Outperforms Most Study Apps

In 1885, Hermann Ebbinghaus memorized thousands of nonsense syllables, then timed exactly how fast he forgot them. His forgetting curve — showing that people lose roughly 50% of new information within an hour and up to 90% within a week — remains one of the most replicated findings in cognitive psychology. More importantly, Ebbinghaus also discovered the correction: review material at the precise moment it is about to slip away, and memory consolidates far more durably than massed study ever produces. That insight became the engine of spaced repetition.

Spaced repetition is a learning technique in which review sessions are scheduled at expanding intervals timed to coincide with the natural decay of memory traces. Rather than rereading notes nightly, a learner revisits material after one day, then three days, then a week, then a month — stretching the gap each time recall succeeds. Research suggests this schedule can reduce forgetting by 80% compared with equivalent time spent in massed (crammed) study.

Why the Brain Responds to Spaced Practice

Memory consolidation depends on two overlapping processes. The first is storage strength — how securely an engram is encoded in long-term structures such as the hippocampus and neocortex. The second is retrieval strength — how readily that engram can be accessed at any given moment. Robert Bjork at UCLA coined the term desirable difficulty to describe the counterintuitive finding that making retrieval slightly harder at the moment of practice strengthens storage far more than effortless review does.

When a memory has partially faded, the neural effort required to reconstruct it triggers processes associated with memory reconsolidation — essentially a re-encoding event that reinforces the synaptic connections involved. Reviewing too early, while the memory remains vivid, produces minimal reconsolidation benefit. Waiting until the trace is almost gone risks complete loss. Spacing review at the optimal moment — just before forgetting — hits the consolidation sweet spot every time.

The Role of Sleep

Sleep amplifies spaced repetition's benefits. During slow-wave sleep, the hippocampus replays newly acquired memories and transfers them to cortical storage, a process called systems consolidation. Studies using EEG monitoring show that memories reviewed with spaced intervals show stronger cortical reactivation during subsequent sleep than those studied in massed sessions, suggesting the two mechanisms are synergistic rather than independent.

Algorithms That Schedule Forgetting

Modern spaced repetition systems (SRS) automate the scheduling problem Ebbinghaus solved by hand. The most influential algorithm is SM-2, developed by Piotr Wozniak in 1987 for his SuperMemo software. SM-2 assigns each flashcard an ease factor — a multiplier that increases when recall is fast and accurate and decreases when hesitation or errors occur. The interval to the next review is calculated as:

• First successful recall: review after 1 day
• Second successful recall: review after 6 days
• Subsequent recalls: interval × ease factor (typically 2.5 for new cards)
• Failed recall: interval resets to 1 day, ease factor decreases

Anki, the open-source SRS used by hundreds of thousands of medical students worldwide, built its scheduling engine on SM-2 and later introduced the FSRS (Free Spaced Repetition Scheduler) algorithm, which uses a machine-learning model trained on 1.7 billion review logs to predict each individual's forgetting curve more precisely than the fixed ease-factor approach.

Empirical Results Across Domains

Practical Implementation

Spaced repetition is most effective when certain conditions are met. Card design matters enormously — asking What causes synaptic strengthening during memory reconsolidation? forces retrieval; simply rereading a definition does not. Each card should test a single, atomic fact. Mixing question types (image-to-definition, fill-in-the-blank, reverse recall) distributes learning across multiple retrieval pathways.

• Daily minimums beat marathon sessions: twenty minutes every day outperforms three hours on weekends in long-term retention metrics
• New cards vs. reviews ratio: most SRS users add 10–20 new cards daily and reserve the remainder of session time for due reviews
• Retiring mastered cards: cards exceeding a six-month interval can often be suspended; retrieval at that point is near-automatic
• Image occlusion: for visual material — anatomical diagrams, geographical maps — hiding regions of an image is more effective than text prompts alone

Limits and Misapplications

Spaced repetition excels at factual recall but is less suited to procedural skills that require physical repetition (e.g., playing an instrument) or conceptual understanding that benefits from worked examples and problem-solving rather than bare fact retrieval. Loading a deck with poorly understood concepts — sometimes called premature SRS — leads to rote memorization without comprehension. Cognitive scientists recommend thoroughly understanding material before converting it to flashcard form.

Overloading a deck is a common failure mode. When daily review queues balloon to several hundred cards, learners often skip sessions, causing backlog-driven abandonment. Research on adherence suggests smaller, well-curated decks with consistent daily engagement produce better long-term outcomes than large decks maintained inconsistently.

Spaced Repetition in Institutional Settings

The Forgetting Curve as an Asset

The counterintuitive implication of spaced repetition research is that forgetting is not simply the enemy of learning — it is part of the mechanism. Each act of retrieval after a period of forgetting rewrites the memory trace with greater fidelity. Deliberately allowing partial forgetting, then recovering the information, is the motor behind durable long-term memory. Ebbinghaus mapped the problem in 1885. The algorithms running on millions of devices today are still, at their core, his answer.