Deliberate practice
Doing something for a long time does not, on its own, make a person better at it; once feedback stops exposing your limits, you tend to plateau. What builds skill is a narrower kind of effort: working at the edge of what you can currently do, with a clear goal aimed at a specific weakness, immediate feedback, repetition with adjustment, and full attention. By design it is uncomfortable — practice that feels easy and pleasant is not this kind of practice.
K. Anders Ericsson's name for the specific kind of practice that actually builds skill. The claim is that ordinary experience — doing the thing a lot, even for decades — does not on its own make a person better at it. What does is a narrower activity with four conditions: a clear goal aimed at a specific weakness, immediate feedback on whether the effort hit the goal, repetition with adjustment, and full attention. The effort is, by design, uncomfortable; if practice feels easy and enjoyable, it is not deliberate practice in the technical sense.
Etymology§
The term and its full formulation come from Ericsson, Krampe and Tesch-Römer's 1993 paper in Psychological Review — the empirical basis was a study of violinists at the West Berlin Hochschule der Künste, comparing the accumulated practice hours of students rated best, good, and teacher-track by their professors. The paper made deliberate practice a technical term and the 10,000-hour figure (the mean practice total of the best group by age 20) one of its incidental findings. Malcolm Gladwell's Outliers (2008) turned that mean into a popular threshold rule; Ericsson disowned the framing publicly and repeatedly — there was nothing magic about 10,000 hours, and the figure described one elite population, not a general law.
Most people who have done a job for twenty years are not twenty years better at it than someone who has done it for two; they are roughly as good as they were when they stopped getting feedback that exposed their limits. Time on task, in Ericsson's argument, does not produce skill. What produces skill is the specific activity of working at the edge of present ability with feedback that tells you whether the next attempt was better than the last.
The four conditions in the 1993 paper are the operational definition. A clear, narrow goal — play this passage without rushing the third bar — rather than a general aim. Feedback fast enough and specific enough that the practitioner can tell whether the goal was met on a given attempt. Repetition with deliberate adjustment of what was off. And full attention through all of it, which is why deliberate practice cannot be sustained indefinitely — Ericsson's elite performers practised in this mode for roughly four hours a day, in 60–90-minute blocks, and rested between.
A common misreading is that any practice counts as deliberate practice as long as the practitioner is focused. The 1993 paper is specific that the feedback condition is structural, not motivational — most domains require either a coach, a marked-up score, or an instrument that closes the loop directly (a chess engine, a metronome, a video replay). Without that, the practitioner cannot tell which attempts were closer to the goal, and the activity reverts to ordinary repetition.
The Berlin violinist study and several follow-up studies in music, chess, and sport showed strong correlations between accumulated deliberate practice and performance level. The meta-analyses that came later — Macnamara and colleagues, 2014 and 2016 — argued that deliberate practice accounts for a smaller share of the variance in performance than the original work suggested: roughly a quarter of variance across domains, and considerably less in domains where the task itself is less stable (business, medicine, research). The Transfer on Trial strand argues that even skilled performance acquired through deliberate practice transfers narrowly — what was built in one task does not generalise to a near neighbour without further deliberate work in the new setting.
The applied form of the construct is the question of how to practise a skill that has no obvious coach, no marked-up score, and no instrument that closes the loop. Research, writing, and working with AI — the three skill-formation worked examples — all sit in that awkward zone: practitioners can work at them for years without identifiable feedback, and most do. The deliberate-practice frame turns the question of how to get better at writing into the narrower question of what is the specific weakness, and what is the source of immediate feedback that exposes it.