Over-learn until retrieval is fluent, then stop

Continue practice just past the point of first correct retrieval — over-learning compresses the forgetting curve without requiring proportionally more time.

Why it works

Over-learning — continuing practice beyond the first correct recall — extends the consolidation of a memory trace. Research shows that a modest amount of over-learning (roughly 50% additional practice beyond first correct recall) produces disproportionate retention gains on delayed tests, because the additional retrieval events solidify the trace before natural decay begins. However, the returns diminish sharply: doubling the over-learning time does not double retention, so extended over-learning is less efficient than spaced review.

How to do it

1. When you can first recall an item correctly, do not stop — attempt it again 2–3 more times to ensure the retrieval is fluent.
2. Stop adding practice when retrieval is consistent and fast, not when it is just barely correct.
3. Invest the saved time in spaced reviews of other material rather than extended over-learning of this item.
4. Apply over-learning selectively to the highest-value items, not uniformly to everything.

Evidence

Krueger (1929) is the classic study on over-learning, showing that groups who practiced 50% beyond first mastery retained significantly more on delayed tests. The diminishing returns above 50% have been documented in follow-up research. (observational)

Krueger’s work used verbal lists; generalization to complex skills has partial support but is less precisely quantified. The "50% more" figure is approximate and varies by material.

Sources

• Krueger, W.C.F. (1929), The effect of overlearning on retention, Journal of Experimental Psychology

Common mistake

Either stopping at first correct recall (leaving retention fragile) or continuing practice far past fluency (diminishing returns), when the productive range is the narrow band just past first mastery.