Task time falls as a power function of practice trials
Aliases: learning curve · skill acquisition
What it is
The power law of practice holds that time for repeated performance of a comparable task often declines as a power function of practice trials: operation starts slow, improves rapidly, then gains taper. It describes skill acquisition under stable conditions, not a guarantee that every person or flow has one identical curve.
Why it happens
Practice reduces search and deliberation, builds efficient movement sequences, and strengthens retrieval cues and expectations. Early trials remove obvious hesitation and error; later performance is limited by finer motor, perceptual, and coordination improvements, making a power function more plausible than a fixed time reduction per trial.
Studying it
Have the same participants repeat a clearly specified, stable task, recording time, errors, strategy, and interruption on every trial. Fit a learning curve on a log scale and report individual patterns. Onboarding comprehension, post-break forgetting, and system delay should not be mistaken for pure skill change.
Where it stops holding
One curve fails when tasks change, gaps are long, feedback is unstable, or strategy differs. A grand mean can hide fast learners and persistent failures. Faster time is not automatically better quality; it can coincide with missed checks or riskier action.
Applying it
- Keep steps, locations, and feedback stable for frequent flows so repetition accumulates rather than requiring relearning.
- Test first success, short-run repetition, and return after a delay separately rather than deciding from an expert demonstration.
- Track errors and outcome quality with time; if speed rises while essential checking disappears, revise feedback or task constraints.