Peak

Ericsson's construct is narrow and specific: practice designed by someone who knows the field, targeting a well-defined weakness just beyond current ability, with immediate informative feedback, and sustained full concentration.

Almost nothing people call practice meets this definition. Playing through a piece, re-reading notes, or doing the job are not deliberate practice, and his research finds they produce little improvement past a plateau.

Most practice is repetition of what you can already do. That is not what the research measured.

Devmark’s reading — not a quotation from the book

He is direct that the number came from Gladwell's popularisation of one study, that it was an average across violinists at one institution with enormous individual variation, and that it appeared in no recommendation of his.

His objection is not pedantry. The framing suggests that time is the variable, which is precisely the opposite of the finding — the quality and structure of the practice dominate, and time spent badly produces nothing.

His theoretical contribution, and the part most often omitted from summaries. Experts differ from novices chiefly in having richer, more detailed mental representations of their domain.

A chess master does not calculate faster; they perceive the board in larger meaningful chunks. Deliberate practice works, in his account, by building these representations — which is why practice that does not challenge them does not improve performance.

A distinction he introduces for fields without an established pedagogy. Purposeful practice has specific goals, focus, feedback and discomfort but lacks an expert-designed curriculum; deliberate practice adds that curriculum.

This matters practically because many domains — management, most knowledge work — have no established training method, so purposeful practice is the best available and the book is honest that it is a weaker instrument.

Ericsson's position is that innate talent plays a much smaller role than assumed, and that most apparent prodigies had unusually early, intensive, well-structured training.

This is his most contested claim. Meta-analyses have found that deliberate practice explains a substantial but far from complete share of performance variance — more in highly structured domains like chess and music, much less in less structured ones — and Ericsson disputed those analyses in turn.

Practice explains a large share of variance in structured domains and much less elsewhere. That dispute is unresolved.

Devmark’s reading — not a quotation from the book

His practical guidance: find a teacher who can design the practice, isolate specific weaknesses rather than practising generally, seek immediate feedback, expect the work to be unpleasant, and maintain full concentration for shorter periods rather than partial attention for longer ones.

The concentration point is under-appreciated — his research subjects practised in limited sessions because full-intensity practice cannot be sustained for long.

His account of why improvement stops: performance becomes automatic, and automaticity is the end of learning. The remedy is deliberately disrupting the automatic pattern by targeting a component just beyond current capability.

This explains the common experience of decades of experience producing no improvement — a well-documented finding in medicine and other professional fields.

Their requirement that is hardest to satisfy: deliberate practice as they define it depends on someone who already knows what good looks like, can see what you are doing wrong, and can design the next exercise for you.

Their treatment of how to choose a teacher, and of what to do when no qualified one exists, is the most practical chapter and the one most often skipped. Their fallback — study the best performers closely enough to build your own model of correct execution, then find any source of honest feedback — is weaker than a coach and considerably better than nothing.

Deliberate practice assumes someone who can already see what you are doing wrong. That is the binding constraint.

Devmark’s reading — not a quotation from the book

The core finding — that structured, feedback-rich, effortful practice outperforms unstructured repetition — is well supported and not seriously disputed.

The stronger claim about talent's minimal role is disputed, and the evidence suggests domain matters considerably. A reader should take the practice methodology confidently and the talent argument as one side of a live disagreement.

The definition itself is the value: check whether what you call practice has a specific target, immediate feedback, and discomfort. If it has none of these, it is repetition and will not move you.

And the plateau diagnosis, which explains why experience alone so frequently produces no improvement at all.