The Forgetting Curve in Elementary Math

Ebbinghaus in One Paragraph

In 1885, Hermann Ebbinghaus tested himself on lists of nonsense syllables. He logged how much he could recall at different time intervals. The result was a steep early drop followed by a long tail: about 50 percent gone in an hour, 70 percent gone in a day, close to 90 percent gone in a week. The curve has been replicated in dozens of studies on real content (vocabulary, medical facts, paired associates, and yes, math) and the shape does not change.

Why Math Facts Hit the Curve Hard

Math facts have two properties that make them especially vulnerable to forgetting:

• They are arbitrary

  There is no internal logic to 7 × 8 = 56. The brain cannot reconstruct it from first principles the way it can re-derive a spelling rule.

• They are atomic

  Each fact is independent. Forgetting one does not flag the others, so the curve drops fact by fact without warning.

• They are rarely used in isolation

  By the time a kid hits 7 × 8 inside a word problem, the working memory is already loaded with the problem context. A slow lookup costs the whole problem, not just the fact.

Estimated Retention Without Spaced Review

For a math fact a kid learned once today, the rough retention shape (based on the standard Ebbinghaus curve) is:

• 1 hour

  ~50% remembered

• 24 hours

  ~30% remembered

• 3 days

  ~20% remembered

• 1 week

  ~10% remembered

• 1 month

  ~5% remembered (effectively forgotten)

Cramming vs Spaced Review: A Side-by-Side

The contrast is so stark it is worth putting one column next to the other. Same kid, same facts, same total practice minutes — different distribution.

Cramming (one long session, no review)

• End of session

  ~100% recall. Feels like total mastery.

• Day 1 (next morning)

  ~50% retained.

• Day 3

  ~20% retained.

• Week 1

  ~10% retained.

• Month 1

  ~5% retained. The session might as well not have happened.

Spaced review (5 minutes/day, SRS schedule)

• End of first session

  ~80% recall. Slightly lower than cramming on day 0.

• Day 1 (after review #1)

  ~85% retained.

• Day 3 (after review #2)

  ~90% retained.

• Week 1 (after review #3)

  ~92% retained.

• Month 1 (after reviews #4 and #5)

  ~90% retained. Almost all of it still there.

Cramming wins the first 60 minutes and loses every interval after that. Spaced practice gives up a little on day 0 and holds 18x more memory by the end of the month. That gap is the entire reason a 5-minute daily routine outperforms a 30-minute weekly worksheet.

What Each Review Buys You

One successful retrieval at the right interval roughly doubles how long the fact survives. With three reviews timed correctly, the same fact that would have been 5 percent retained at 1 month is over 90 percent retained.

• Review at 24 hours

  Retention at 1 week jumps from ~10% to ~60%

• Review at 3 days

  Retention at 1 month jumps from ~5% to ~70%

• Review at 7 days

  Retention at 1 month climbs to ~85%

• Review at 14 days

  Retention at 3 months climbs to ~90%

That is the entire spaced repetition schedule, derived directly from the curve. Read spaced repetition for math facts for the implementation details.

The Summer Slide Connection

Summer slide is the same curve over a longer interval. Two and a half months with zero retrieval pushes every fact deep into the tail of the curve. The kid does not lose all their math — the durable, automatic stuff is fine — but the fluent-but-not-yet- automatic facts disappear. That is why September is full of review weeks.

A 5-minute daily session through the summer keeps the curve flat. No new content; just enough retrieval to prevent the slide. See 5-minute math practice for the routine.

Put the Curve to Work

Spaced repetition is not a productivity hack. It is the engineered response to a memory curve we have measured for over a century. For the comparison with traditional drills see spaced repetition vs traditional math drills. For the Anki / SM-2 framing see Anki-style multiplication practice, or jump straight to the under-3-second method in how to build math automaticity. To start a session, open practice.