Check for diminishing returns before continuing

Beyond roughly 50–100% overlearning, additional practice yields less return per hour.

Why it works

The overlearning curve flattens steeply. Early overlearning trials (up to 50% of baseline) produce the largest gains in long-term retention. Trials between 50–100% show smaller but still positive effects. Beyond that, the marginal gain per practice trial drops toward zero while the opportunity cost — time not spent on the next skill — grows. The check prevents overlearning from becoming perfectionism that crowds out progress.

How to do it

1. After 50% overlearning, test retention at a delay (at least 24 hours) to see whether it has actually improved.
2. Compare the value of additional drilling to the value of beginning a new skill.
3. If performance is consistent at delay, shift to spaced maintenance reviews rather than continued massed drilling.
4. Reserve maximal overlearning for skills that will be tested under extreme conditions; use lighter overlearning for everyday competencies.

Evidence

Driskell et al. (1992) meta-analysis found clear benefit from overlearning at 50% and 100% of criterion trials for retention, with returns diminishing beyond that range. Excess overlearning beyond 200% showed little additional benefit. (observational)

These estimates are averages across tasks; the optimum varies by skill complexity and retention interval. Use them as heuristics, not precise prescriptions.

Sources

• Driskell, Willis & Copper (1992), effect of overlearning on retention, Journal of Applied Psychology

Common mistake

Continuing to drill a skill that already tests perfectly, mistaking the comfort of familiar repetition for productive practice, while neglecting skills that actually need work.