Deliberate practice

Why do some people keep improving at a craft for decades while others stop improving after a year?

Someone who has driven for thirty years is not, on average, a better driver than someone who has driven for five, and someone who has typed daily since 1998 does not type more accurately than they did in 2001. Yet a violinist of forty years' standing is unmistakably better than she was at twenty. All three accumulated hours; only one accumulated skill.

So "practice makes perfect" is hiding something. It quietly assumes that repeating an activity is the same as working on it. Is it?

Ask what a change in performance would require. To do something better tomorrow than today, three things seem necessary: an attempt at something you cannot yet do reliably; information about how the attempt fell short; and a chance to alter the next attempt in light of it. Take away any one — what is left?

Take away the first. The typist stopped attempting anything difficult years ago. Skill acquisition runs from effortful and attention-hungry to smooth and automatic, and automaticity is exactly what makes a skill useful in daily life. But notice the cost: once the task runs itself, you stop producing a gap between what you intended and what you did. No gap, no signal, nothing to change. The plateau is less a ceiling on talent than the moment the task stopped being hard.

Take away the second instead. Imagine hitting golf balls into thick fog. A thousand effortful swings teach nothing, because nothing tells you where the ball went — and worse, you become fluent at whatever you happened to be doing, errors included. Does that suggest why some fields produce experts and others produce confident veterans? Where feedback is fast, clean and attributable — a chess move, a suture, a bar of music — experience compounds. Where it arrives years later, tangled with luck and other people's decisions, decades of it can leave judgement almost untouched.

So the practice that keeps working has a shape: tasks pitched just past current reach, feedback immediate and specific enough to act on, repetition with correction, and usually someone outside you, because you cannot hear the errors you do not yet have the ears for. It is effortful and not much fun, which explains its rarity better than any theory of willpower.

One honesty note, because the idea is often oversold. How much of the gap between experts it accounts for is genuinely contested: large re-analyses find accumulated deliberate practice explains a substantial minority of the variation in music, games and sport, and strikingly little in professional work, with starting age, physical constraints, opportunity and feedback quality taking up much of the rest. It looks necessary for sustained improvement; it was never sufficient for eminence.

Two corrections. The famous "ten thousand hours" is a popularization: the underlying study reported an average accumulated by one group of elite students, not a threshold anyone must cross or a guarantee for anyone who does. And deliberate practice is not a synonym for practising hard — a punishing two-hour session doing what you can already do is repetition; twenty attentive minutes on the passage you keep fumbling, with the metronome telling you the truth, is not.

• Why domains with slow, noisy feedback — forecasting, management, therapy — produce confidence far faster than accuracy.
• How a good teacher's main function may be diagnostic: seeing the error the learner cannot yet perceive.
• Whether the same logic explains organisational learning, where the feedback loop runs through many hands.

Every term the collection defines is gathered in the glossary.