Expertise is built through years of focused effort, but practice alone accounts for far less of expert performance than most people assume. A large meta-analysis across multiple domains found that deliberate practice explained only about 1% to 26% of the difference in performance between individuals, depending on the field. What fills the rest of that gap involves genetics, how the brain physically reorganizes itself, the quality of feedback a learner receives, and forms of knowledge that experts themselves struggle to articulate. The science of expertise has moved well past the “10,000-hour rule” into messier, more interesting territory.
How Much Practice Actually Matters
The popular idea that enough practice can make anyone an expert traces largely to research on violinists published in 1993 by K. Anders Ericsson and colleagues. That original study reported a tight correspondence between accumulated practice hours and skill level among elite performers. But when a team of researchers attempted to replicate those findings using more rigorous methods, the relationship weakened considerably. Controlling for statistical biases and error inflation, the amount of deliberate practice explained substantially less variation in violin performance, and among the most elite performers, practice hours could not account for why some reached higher levels than others.1PubMed Central. The role of deliberate practice in expert performance: revisiting Ericsson, Krampe & Tesch-Römer (1993)
A broader meta-analysis covering music, games, sports, education, and professional work put clearer numbers on the question. Deliberate practice explained about 26% of the variance in game performance, roughly a fifth for music, 18% for sports, 4% for education, and less than 1% for professional performance.2PubMed. Deliberate practice and performance in music, games, sports, education, and professions: a meta-analysis That last number is striking: in real-world professional settings, the sheer volume of practice barely registers as a predictor of who performs best.
This does not mean practice is useless. It means the type and quality of practice matter as much as, if not more than, the total hours logged. Researchers who critiqued those meta-analyses argued that many of the studies they included measured generic “structured practice” rather than truly individualized, teacher-guided deliberate practice as Ericsson originally defined it. In a study of chess players, when practice was highly individualized and guided by quality instruction, the predictive power of practice hours was more than three times higher than what the meta-analyses reported at the average level.3Current Psychology. The meta-analyses of deliberate practice underestimate the effect size because they neglect the core characteristic of individualization So the gap between “I practiced a lot” and “I practiced the right things with expert guidance” turns out to be enormous.
The Genetic Side of the Equation
One reason practice alone falls short is that genetic differences influence both baseline ability and the rate at which people improve. A study of twins raised apart found that the heritability of motor skill performance was high even before any practice began. More surprising, the rate of learning itself was heritable. As practice accumulated, genetic influences on performance actually grew stronger rather than weaker, because practice reduced the noise introduced by different environmental backgrounds.4PubMed. Genetic and environmental contributions to the acquisition of a motor skill In other words, the more everyone practiced, the more innate differences stood out.
Genetic influence extends beyond motor skills into cognitive domains. Research on musical expertise found that polygenic scores for cognitive ability predicted not only general intelligence but also musical achievement and auditory discrimination. People with higher genetic predisposition for cognitive performance got more out of each hour of practice: the association between practice and musical expertise grew stronger as genetic scores went up.5Heliyon. Gene-environment interaction in expertise acquisition: Practice effects on musical expertise vary by polygenic scores for cognitive performance Genetics and practice are not competing explanations. They interact, with each amplifying the other.
None of this implies that talent is fixed or that effort is pointless. What it does mean is that two people investing identical hours of high-quality practice will not arrive at the same place. The playing field was never level, and that is worth acknowledging honestly rather than pretending otherwise with motivational slogans about grit.
How Expertise Reshapes the Brain
When someone trains intensively in a skill, their brain does not simply “get better” at using the same circuits. It physically reorganizes how and where it processes information. Brain imaging studies consistently show that as people acquire skills, activity in areas associated with effortful control, including the prefrontal cortex and anterior cingulate cortex, reliably decreases. Meanwhile, activity in subcortical regions involved in automatic processing increases.6PubMed. Functional brain changes following cognitive and motor skills training: a quantitative meta-analysis The brain essentially moves the work from the slow, deliberate “thinking hard” circuits to faster, more efficient ones.
This pattern is sometimes called neural efficiency, and it shows up across very different domains. When expert archers were scanned during aiming, their brain activity was more localized and efficient compared to novices, who activated broader and less focused neural networks.7Cognitive and Behavioral Neurology. An fMRI Study of Differences in Brain Activity Among Elite, Expert, and Novice Archers at the Moment of Optimal Aiming A study of elite athletes from multiple sports found that when athletes imagined movements from their own sport, they showed superior behavioral performance with lower levels of brain activation compared to imagining movements from unfamiliar sports, confirming that expertise allows better output with less neural energy.8PubMed Central. Neural Efficiency and Acquired Motor Skills: An fMRI Study of Expert Athletes
One cognitive framework proposes that experts achieve this efficiency by building “chunks” and larger knowledge structures that allow portions of long-term memory to function as an extension of working memory. This reorganization follows a two-stage process: first a decrease in brain activation as basic patterns become familiar, then a functional reorganization as those patterns get integrated into larger, more flexible structures.9PubMed. How chunks, long-term working memory and templates offer a cognitive explanation for neuroimaging data on expertise acquisition: a two-stage framework When a chess master glances at a board and immediately grasps the strategic situation, they are not thinking faster than a beginner. They are recognizing patterns that the beginner has to construct piece by piece.
How Experts See and Decide Differently
Expert perception is qualitatively different from novice perception, not just faster or more accurate. Eye-tracking research in radiology reveals that expert radiologists use a distinctive “global-focal” search pattern: they take in the whole image first, forming a rapid global impression, then shift into a detailed focal search. Novices, by contrast, scan more haphazardly. Interestingly, one study that tried teaching novices to adopt expert-like search strategies found no significant improvement in their perceptual performance, suggesting these patterns emerge from deep familiarity rather than surface technique.10PubMed Central. How visual search relates to visual diagnostic performance: a narrative systematic review of eye-tracking research in radiology
A similar gap appears in cardiology. Expert cardiologists reviewing cardiac imaging spent less time overall, made fewer eye movements, and were less variable in their review times compared to novices. Novices fixated repeatedly on critical frames and regions but this did not help them reach better diagnoses.11PubMed Central. Eye-tracking for assessing medical image interpretation: A pilot feasibility study comparing novice vs expert cardiologists The expert does not see the same image more carefully; they see a fundamentally different image because their visual system has been trained to extract the diagnostically relevant information almost automatically.
In motor-skill domains, a parallel phenomenon called “quiet eye” shows up. This refers to a final, steady fixation on a target before executing a movement. Expert air pistol shooters hold this fixation for roughly 925 milliseconds on average, compared to about 330 milliseconds for novices.12European Journal of Physical Education and Sport Science. Comparison of Quiet Eye Duration of Three Different Performance Level Athletes in Air Pistol Shooting Training golfers to extend their quiet eye period improved their putting performance under pressure, suggesting this gaze behavior is not merely a marker of expertise but part of its mechanism.13PubMed Central. The effect of quiet eye training on golf putting performance in pressure situation
Expert decision-making also relies heavily on intuition, but only in the right environments. Research found that intuitive judgments from experts are trustworthy when two conditions are met: the environment is regular enough that its patterns can be learned, and the person has had sufficient opportunity to learn those patterns through repeated feedback.14PubMed. Conditions for intuitive expertise: a failure to disagree An experienced firefighter’s gut feeling about when a floor is about to collapse can be genuinely reliable. A stock picker’s gut feeling about next quarter’s market probably is not, because financial markets do not provide the same kind of stable, learnable feedback.
When Expertise Backfires
Deep expertise creates real vulnerabilities alongside its advantages. The most well-studied of these is the Einstellung effect, in which a familiar solution that comes to mind first actively prevents an expert from finding a better one. In chess experiments, expert players who had already identified a workable solution reported that they were searching for something superior. But eye-tracking data told a different story: their gaze kept drifting back to features of the board related to the first solution they had found, even when a clearly better move existed elsewhere.15PubMed. Why good thoughts block better ones: the mechanism of the pernicious Einstellung (set) effect The experts were not being lazy; their own pattern-recognition machinery was steering their attention away from the better answer.
Eye-tracking work on anagram solving found a related dynamic. When the central letter string formed a familiar word, participants encoded it faster but then struggled more to rearrange the letters into the correct solution. Familiarity helped with encoding but interfered with flexible recombination.16PubMed Central. The Einstellung effect in anagram problem solving: evidence from eye movements The same strength that makes experts fast, their rich network of learned patterns, can also make them rigid.
Experts are also vulnerable to choking under pressure, and the mechanisms differ from how novices fail. Leading theories propose that high-stakes situations can cause experts to shift attention back to the step-by-step details of a skill they normally execute automatically, essentially overriding the efficient processing they spent years developing. Alternatively, pressure may distract attention away from the task or simply elevate arousal past the optimal range.17PubMed Central. Choking under pressure: the neuropsychological mechanisms of incentive-induced performance decrements Either way, the result is the same: the expert temporarily loses access to the very automaticity that defines their skill.
Does Expertise Transfer Across Domains?
A persistent hope in education and training is that getting really good at one thing will make you better at related things. The evidence is discouraging. A review of research on chess training, music training, and working-memory training found that while these activities produced small to moderate improvements on measures beyond the trained task, the effect sizes shrank as study designs got more rigorous. Once proper control groups and blinding were in place, far transfer of learning rarely occurred.18PubMed Central. Does Far Transfer Exist? Negative Evidence From Chess, Music, and Working Memory Training
There are exceptions, but they tend to be modest. Research on cognitive training in older adults has found some evidence of far transfer when training involves teaching explicit strategies rather than just repeated practice.19PubMed Central. Far transfer in cognitive training of older adults The distinction matters: brute repetition seems to build narrow skills, while learning transferable strategies can sometimes produce broader benefits. But “sometimes” and “modest” are the operative words. If you learn chess hoping it will make you better at math, prepare to be disappointed. You will mainly get better at chess.
What Experts Know Without Knowing They Know It
A substantial portion of expert knowledge is tacit, meaning experts cannot fully articulate what they know or how they make decisions. This tacit knowledge develops through implicit learning, the largely unconscious absorption of statistical patterns and regularities in the environment. Researchers have argued that this implicit learning acts as a form of scaffolding for expertise, supporting the acquisition, retention, and transfer of skill in ways the learner is never fully aware of.20Journal of Cognitive Engineering and Decision Making. Implicit Learning, Tacit Knowledge, Expertise Development, and Naturalistic Decision Making
Expert musicians offer a vivid example. Brain recordings show that trained pianists produce a distinct neural signal roughly 100 milliseconds before they physically press a wrong key. Their brains have already detected the error before the finger completes its movement.21PLoS ONE. Nobody Is Perfect: ERP Effects Prior to Performance Errors in Musicians Indicate Fast Monitoring Processes A separate study found that pianists with early and extensive training showed a specific brain-wave component that was entirely absent in untrained players, reflecting a mismatch between the sound they intended to produce and the sound they actually heard.22PubMed. Feedback-based error monitoring processes during musical performance: an ERP study These lightning-fast monitoring processes are not conscious. The expert’s brain is running a predictive model of what should happen next and flagging deviations before the conscious mind even registers them.
Tacit knowledge creates a real problem for training and feedback. If the expert cannot explain what they are doing differently, standard instruction becomes harder. Early feedback during training appears to help: in a study of surgical trainees learning an endoscopic procedure, those who received expert feedback early in a spaced-practice schedule performed faster and scored higher than those who received the same feedback later.23Simulation in Healthcare. Optimizing the Timing of Expert Feedback During Simulation-Based Spaced Practice of Endourologic Skills Timing matters because early feedback can shape how learners encode a skill from the start, before bad habits become entrenched. But even the best feedback system cannot fully externalize what an expert knows tacitly.
Maintaining Expertise as You Age
General cognitive and physical abilities decline with age, yet experts in many fields sustain high performance long after those declines have set in. Research identifies three explanations for this. First, the specialized abilities experts rely on may be partially preserved from age-related decline because of years of training, a concept called preserved differentiation. Second, experts develop compensatory strategies that work around declining capacities. Third, experts selectively maintain the specific abilities most critical to their performance while letting less important ones fade.24European Review of Aging and Physical Activity. Expertise and aging: maintaining skills through the lifespan
Continued deliberate practice is key to maintaining performance in old age. Much of what looks like inevitable age-related decline in skilled performance is actually attributable to reduced engagement in domain-related activities. Older experts who keep practicing maintain their level, while those who scale back see performance drop in ways that track their reduced practice more than their biological aging.25Journal of Aging and Physical Activity. How Experts Attain and Maintain Superior Performance: Implications for the Enhancement of Skilled Performance in Older Individuals An aging surgeon who keeps operating may retain their skills far longer than a slightly younger one who shifted to administrative work.
Expertise in Teams and Machines
Individual expertise does not automatically produce group expertise. Assembling a team of highly skilled individuals is not the same as building an expert team. Research on team performance identifies principles that distinguish the two, including shared mental models, role clarity, psychological safety, and dynamic leadership. A group of experts who cannot coordinate, share information effectively, or adapt their roles to shifting demands will often underperform a well-coordinated group of less individually skilled people.
Artificial intelligence has added a new dimension to the expertise question. In a study comparing human clinicians and an AI model at detecting a specific wound condition, the AI achieved about 90% accuracy while humans averaged roughly 79%. But the gap was not uniform across all humans. When tested for statistical significance, the AI only meaningfully outperformed those clinicians who lacked relevant formal qualifications, had low diagnostic confidence, did not focus on wound care, or had short work experience. Experienced clinicians with relevant training performed comparably to the AI.26Journal of the American Medical Informatics Association. Diagnostic accuracy differences in detecting wound maceration between humans and artificial intelligence: the role of human expertise revisited
That finding captures something important about the current state of AI and expertise. Machine learning systems excel at pattern recognition in well-defined, data-rich environments, which happens to be the same territory where human expertise is strongest. Where AI currently struggles is in the less structured, judgment-heavy, context-dependent situations where tacit knowledge and adaptive reasoning matter most. The practical upshot for now is not that AI replaces expertise but that it raises the floor: tasks that once required moderate skill can be automated, while the distinctly human contributions of expert judgment become more, not less, valuable at the margins.
Who Counts as an Expert
Outside the laboratory, “expert” is a social label as much as a cognitive one. Scholars in social epistemology have framed experts as individuals deeply immersed in specialist practices whose claims can be evaluated based on their track record of truth-tracing, their adherence to the standards of their field, and the transparency of their reasoning. Expertise in this view is tentative and context-dependent: a person is an expert relative to a domain and an audience, not in some absolute sense. The challenge for non-experts is recognizing who to trust, which depends on signals like institutional affiliation, peer recognition, and a culture of accountability within the expert’s community.
This social dimension is worth thinking about because the cognitive science of expertise and the public meaning of expertise sometimes diverge. A person with 20 years of experience may have built rich, efficient, largely tacit knowledge that genuinely serves them well. Or they may have spent 20 years repeating the same comfortable routines without ever getting meaningful feedback, in which case their experience confers confidence far more than competence. The research consistently shows that raw experience is a poor proxy for expertise. What separates genuine experts from experienced non-experts is not time served but the quality of their practice, the environments they learned in, and whether they have continued to encounter and adapt to new challenges.