The standard narrative of professional aging is a simple curve. Skills accumulate through midlife, plateau, and then erode as processing speed and working memory decline. It is an intuitive story, and it is substantially wrong. Decades of lifespan research show that expertise does not age as a single capacity. Some domains show remarkable preservation into the seventh and eighth decades. Others show measurable decline well before practitioners themselves notice it.

The puzzle is that the underlying cognitive architecture ages in broadly similar ways across individuals. Fluid abilities such as speed, working memory capacity, and inhibitory control follow a fairly reliable downward trajectory from early adulthood. Yet a 70-year-old concert pianist may execute passages at speeds indistinguishable from younger colleagues, while a physician two decades out of training may follow current evidence less closely than a recent graduate. The brain-aging processes are the same, but the outcomes differ radically.

Resolving this puzzle means setting aside the question of whether expertise declines and asking instead under what conditions it is maintained. Three factors emerge consistently from longitudinal and cross-sectional research: the structure of the domain itself, the persistence of deliberate practice, and the capacity of experts to reorganize how they deploy knowledge. Together, they suggest that expertise is less a possession than an ongoing developmental achievement, one that must be actively sustained against the current of biological aging.

Chess offers one of the most quantified windows into expertise across the lifespan, because rating systems track performance over entire careers. Analyses by Neil Charness and colleagues indicate that chess skill typically peaks in the mid-thirties to early forties and then declines gradually. That decline is shallow relative to the steep drops in processing speed measured in laboratory tasks. Older masters lose ground mainly in time-pressured play and deep calculation. Positional judgment remains comparatively robust.

Aviation research tells a similar but more conditional story. In longitudinal studies of licensed pilots at Stanford and the VA Palo Alto, Joy Taylor, Jerome Yesavage and colleagues found that older pilots performed worse on flight simulator tasks than younger ones. Higher levels of aviation expertise, however, were associated with slower age-related decline over time. Expertise did not abolish aging; it bent the trajectory. The benefit was most evident on domain-relevant demands such as executing air traffic control instructions.

Medicine presents the most sobering picture. A systematic review by Niteesh Choudhry, Robert Fletcher and Stephen Soumerai found that physicians with more years in practice often performed worse on measures of knowledge and adherence to standards of care. Later work by Yusuke Tsugawa and colleagues found that patients of older hospitalists had slightly higher mortality. The exception was physicians who treated high volumes of patients, among whom the age association disappeared. Volume appears to function as a form of continuous calibration.

Music, by contrast, shows striking preservation among those who continue performing seriously. Ralf Krampe and K. Anders Ericsson studied younger and older pianists. Older experts matched younger experts on music-specific speeded tasks, even while showing typical age-related slowing on general measures of processing speed. Older amateurs in the same study showed no comparable insulation. Their performance on musical tasks declined much as their general abilities did.

What distinguishes these trajectories? Domains differ in how much performance depends on raw speed versus accumulated pattern knowledge, how rapidly their knowledge base becomes obsolete, and how much corrective feedback practitioners receive. Medical knowledge evolves quickly, while feedback on diagnostic accuracy is often delayed or absent. A chess player, by contrast, loses rated games immediately. Domains with rich feedback and stable knowledge structures tend to protect their aging experts. Domains with sparse feedback and shifting standards tend to expose them.

Takeaway

Expertise ages according to the shape of the domain. Where feedback is immediate and knowledge is stable, experience compounds. Where feedback is sparse and knowledge shifts, experience can quietly harden into outdated confidence.

The deliberate practice framework, set out by Ericsson, Krampe and Clemens Tesch-Römer in 1993, proposed that elite performance results from sustained, effortful activity designed specifically to improve performance. Such activity is distinct from routine work or enjoyable play. The implication for aging is direct. If expertise is built through deliberate practice, perhaps it is also maintained through it. The later pianist research offered one of the first serious tests of this hypothesis.

Krampe and Ericsson found that among older expert pianists, the amount of deliberate practice undertaken in recent years predicted how well music-specific skills were preserved. Older experts who maintained their skill were not coasting on decades of prior investment. They were still practicing substantially every week. Maintaining a skill appeared to require less practice than acquiring it. It did not, however, require none.

Yet practice patterns typically shift with age in ways that undermine maintenance. Professional work becomes increasingly routinized, and routine performance, however frequent, is not deliberate practice. A surgeon performing a familiar procedure for the thousandth time is exercising automated skills rather than stretching them. So is a radiologist reading routine films. Automaticity is efficient, but it is largely closed to refinement. Years of experience can therefore accumulate without corresponding gains, a pattern Ericsson described as arrested development.

Contextual factors compound this drift. Senior professionals often move into supervisory, administrative or teaching roles that reduce direct performance. Motivational priorities also change. Socioemotional selectivity theory, developed by Laura Carstensen, holds that as time horizons shorten, people prioritize emotionally meaningful goals over acquiring new knowledge. The effortful and often frustrating work of deliberate practice may simply come to feel less worth its cost.

Important limits remain. Deliberate practice preserves domain-specific skill but does little to protect domain-general capacities. Krampe and Ericsson's older experts still slowed on non-musical tasks. Meta-analytic work by Brooke Macnamara, David Hambrick and Frederick Oswald has also questioned how much variance in performance practice actually explains. Their findings suggest that working memory and other basic abilities continue to matter. Practice shields the skill, not the substrate on which it runs.

Takeaway

Experience and practice are not synonyms. Decades of routine performance can coexist with stagnation, while modest but genuinely effortful practice can preserve an expert's edge for decades.

If older experts maintained performance solely by practicing harder, they would eventually lose the race against biological decline. The more interesting finding is that they change how they perform. Timothy Salthouse's classic study of typists showed this vividly. Older skilled typists were slower on basic tapping and choice-reaction tasks, yet typed as quickly as younger typists. They compensated by looking further ahead in the text and preparing keystrokes earlier.

This is selective optimization with compensation in miniature. Paul Baltes illustrated the framework with the pianist Arthur Rubinstein. In late life, Rubinstein reduced his repertoire, which is selection. He practiced the remaining pieces more intensively, which is optimization. He also slowed down before fast passages so they sounded faster by contrast, which is compensation. The output remained virtuosic, but the underlying process had been rebuilt.

Similar restructuring appears in chess. Charness found that older players considered fewer moves than younger players of equivalent skill yet chose moves of comparable quality. They relied more on pattern recognition and evaluative heuristics than on exhaustive calculation. In medicine, Henk Schmidt and Henny Boshuizen described a parallel process called knowledge encapsulation. Detailed biomedical knowledge becomes compressed into higher-order clinical concepts and illness scripts, which allow rapid diagnosis.

Restructuring has costs. Encapsulated knowledge is efficient precisely because it bypasses deliberation, and that also makes it resistant to revision. An expert who relies on well-worn scripts may diagnose quickly and accurately in typical cases. The same expert may struggle with atypical presentations or fail to update when evidence changes. Compensation that depends on pattern libraries is only as good as the patterns, and those libraries can become dated.

Restructuring can also achieve only so much. Ulman Lindenberger, Reinhold Kliegl and Baltes found that professional graphic designers showed the same age differences as non-experts on imagery-based memory tasks, despite the tasks' apparent overlap with their skills. Expertise compensates only within carefully bounded conditions. The most successful aging experts appear to recognize those boundaries. They build environmental supports, checklists and collaborative structures that offload what the aging mind does less well.

Takeaway

The aging expert often succeeds not by doing the same thing equally well but by doing a different thing that produces the same result. Late in a career, the real skill may be knowing what to change.

Taken together, the evidence dissolves the notion of a single curve for how expertise ages. Professional skill follows multiple trajectories, shaped by domain structure, the persistence of effortful practice, and the expert's capacity to reorganize. Decline is common, but within the domain of expertise itself it is neither uniform nor inevitable.

This reframing carries practical weight. Rather than age-based retirement rules, it supports performance-based assessment, structured feedback, and institutional designs that preserve deliberate practice into late career. It also warns against complacency, because experience without feedback is a poor guarantee of skill.

Baltes described development as a lifelong dynamic of gains and losses. Expertise in later life makes that dynamic visible. It is a continuous negotiation between what biology takes and what adaptation can rebuild. The question is rarely whether an expert is aging, but whether they are still developing.