KAELAN.FYI READ / THINK / MAKE
Library
KAELAN.FYI / NOTE 0780 WORDS

understand / seed

Spaced Repetition and Memory

The science of forgetting and how spaced retrieval builds durable knowledge

Seed · early Updated Jul 19, 2026 4 min read
learningmemorypractice

Hermann Ebbinghaus published his memory experiments in 1885. He studied himself as he learned series of nonsense syllables, then measured how much work relearning saved at later times. Forgetting was fast at first and slower later.1 His curve describes that task and person. It is not a fixed rate for every kind of memory.

Spacing study over time often improves later recall. The best gap depends on the material, the learner, and how long the learner needs to retain it. There is no point called “just before forgetting” that a person or app can know with certainty.

This is spaced repetition: reviewing material at expanding intervals to maximize retention with minimum effort.

For delayed retention, retrieval practice can beat spending the same time on more study. The advantage is not universal: repeated study can win on an immediate test, and the result changes with the task. Trying to answer from memory and then checking the answer is one useful form.

The act of retrieval can strengthen later access. Failure can still teach when the learner gets quick corrective feedback. Effort alone is not the goal; a hard guess with no correction can rehearse an error.

In one set of experiments, repeated study won on a test five minutes later, while repeated retrieval won after two days and one week.2 A broad review rated practice testing and distributed practice as high-utility techniques, but their effect still depends on the task and feedback.3

Spaced-repetition software estimates when to show a card again. It uses your answers and a model of memory; it does not observe the memory itself. Anki’s current FSRS scheduler estimates a card’s difficulty and stability from review history and a chosen retention target.4

Regular review keeps overdue work from piling up, but there is no research-backed daily card count for everyone. Add cards at a rate that leaves time to understand them and use the knowledge elsewhere.

What to put on cards:

Factual knowledge that needs to be instant: medical terminology, programming syntax, vocabulary, historical dates, legal standards. Material where slow recall is as useless as no recall.

Conceptual hooks that trigger deeper understanding: the card prompts recall of a concept you understand; you’re maintaining access, not learning the concept from the card.

Material with no natural context for recall: your daily life won’t spontaneously quiz you on the capitals of Southeast Asian countries. Cards provide the context.

What not to put on cards:

Complex conceptual understanding: flashcards work for atoms of knowledge, not essays. You can card “What is the forgetting curve?” not “Explain the theory of spaced repetition.”

Skills: you can’t flashcard your way to speaking a language or playing an instrument. Cards support skills by maintaining vocabulary or factual scaffolding; they don’t replace practice.

Material without context or meaning: if you don’t understand why something matters, memorizing it is fragile. Comprehension first, cards second.

The spacing effect extends beyond flashcards. Interleaving can help learners tell similar problem types or categories apart. It is not a general law that mixed practice beats blocked practice. A meta-analysis found that the benefit changed with the material and the similarity between categories.5

Some tasks that feel harder during practice lead to better later performance. Others are merely hard. Measure the result after a delay rather than trusting either ease or strain.

Anki is one common tool. Its desktop app and AnkiDroid are free; the official iPhone and iPad app is paid.6 Card writing still takes practice.

Effective card design:

  • One fact per card. Atomic knowledge.
  • Clear, unambiguous questions. The answer should be specific.
  • Cloze deletions for connected material. “Ebbinghaus published his memory experiments in {{c1::1885}}.”
  • Images and examples for context. Visual memory is powerful.

The time a deck takes depends on its size, settings, card quality, and missed days. A deck also needs pruning. Delete or suspend cards that no longer serve a real use.

Spaced repetition is maintenance. It does not replace understanding, practice, or checking a high-stakes fact against a trusted source.

Memory is not the whole goal. Spaced retrieval can keep useful facts available, but it cannot decide which facts matter or teach when to use them.

Go Deeper

Books

  • Make It Stick by Brown, Roediger, and McDaniel — Cognitive science of learning, including spacing and testing effects.
  • Fluent Forever by Gabriel Wyner — Spaced repetition applied to language learning.

Essays

  • Michael Nielsen’s “Augmenting Long-term Memory” — Deep dive on personal Anki use for serious learning.
  • Gwern’s “Spaced Repetition for Efficient Learning” — Comprehensive review of research and practice.

Tools

  • Anki — The standard. Requires investment to learn well.
  • RemNote, Mochi — Alternatives with more integrated note-taking.

Sources

  1. Hermann Ebbinghaus, “Memory: A Contribution to Experimental Psychology,” Chapter 7, 1885, translated 1913. ↩

  2. Henry L. Roediger III and Jeffrey D. Karpicke, “Test-Enhanced Learning”, Psychological Science 17, no. 3 (2006), 249–255. ↩

  3. John Dunlosky and colleagues, “Improving Students’ Learning With Effective Learning Techniques”, Psychological Science in the Public Interest 14, no. 1 (2013), 4–58. ↩

  4. Anki Manual, “Deck Options: FSRS”. ↩

  5. Matthias Brunmair and Tobias Richter, “Similarity Matters: A Meta-Analysis of Interleaved Learning and Its Moderators”, Psychological Bulletin 145, no. 11 (2019), 1029–1052. ↩

  6. Anki, “Apps”. ↩