Psychology says simply spending more time on something doesn't guarantee mastery; researchers studying 44,213 chess players found focused training far outperformed repetition

We’ve all heard the “10,000 hours” idea: that enough repetition eventually makes anyone an expert. A new chess study is complicating that story. According to the study, ‘Not All Practice Is Created Equal: Longitudinal Evidence From Over 40,000 Chess Players,’ published in Psychological Science, researchers tracked more than 44,000 Chess. com players over six months and found that how people trained mattered a lot more than how long they played. Players who trained with focused, feedback-driven training improved about 3.6 times faster per hour than players who just kept playing games.

If you’ve ever spent hours doom-scrolling through practice reps, extra sets at the gym, endless flashcards, another round of the same video game, and didn’t feel like you were actually getting better, this research might explain why.

The study that put the "practice makes perfect" myth to the test
The advantage for Daniel A. Southwick of the University of Pennsylvania and his team was rare: real, time-stamped data, not the hazy recollections of how much people practiced. Chess. com told them exactly what each player did: played rated games, reviewed past games move by move, watched short video lessons, solved tactics puzzles, and showed how each player’s skill rating changed over time.


Over 90% of everyone’s time on the platform was spent playing games in real, competitive matches against other players. It’s like the chess version of just showing up and playing pickup basketball for hours. But that wasn’t where the real gains were happening.

Reviewing your mistakes beats playing more games
An hour of "deliberate practice" activities, such as reviewing games, taking lessons, or drilling puzzles, was associated with a 3.11-point rating gain, compared with just 0.86 points for an hour of plain gameplay. That is a 3.61-fold efficiency gap, and it held across the whole sample.

But not all training methods were delivered equally. Reviewing your own games, where an engine flags your mistakes and shows you the better move, was tied to the biggest, most consistent payoff, about 5.6 times more efficient than gameplay. The video lessons scored even higher, but also much more spread out, and less precise. A little surprisingly, puzzles barely edged out gameplay.

The researchers believe the difference comes down to what happens during the activity. When you review a game, you are forced to confront your own particular mistakes and see exactly how to correct them; that’s individualized, targeted feedback. Puzzle apps, especially timed or speed-based ones, can become a “how fast can I solve what I already know” exercise, instead of learning.

Beginners had the most to gain, but even experts benefited
The advantage of focused practice was greatest for newer players and declined somewhat as people got better, a pattern consistent with the natural tapering off of improvement everyone experiences as they approach the edge of their ability. But even for stronger players, going over old games was still a lot more efficient than simply playing more games.

But here’s the catch: it’s a reflection of something psychologists refer to as the “fan-spread effect,” whereby those who get ahead tend to get more ahead over time. It’s often blamed on natural talent. But the current study showed that stronger players spent a much higher percentage of their time in review and lessons than beginners did. That difference in strategy alone, not raw talent, could account for a big chunk of why some learners continue to distinguish themselves from the pack.

What this means beyond the chessboard
The authors are careful to point out real limits here. This is one platform, one skill, and does not capture off-platform practice, coaching, or motivation. It's also a correlational study and not a controlled experiment, and therefore cannot prove that the training choices alone caused the gains in ratings. Most sample players were recreational or developing, not elite competitors, so the findings may not be directly applicable to grandmaster-level training.

Even with those caveats, the takeaway is consistent with decades of research on how skills are actually built. According to the seminal 1993 paper by K. Anders Ericsson and colleagues, published in Psychological Review, expertise tends to come from effortful, structured practice with clear feedback, not just accumulated hours. In the paper, Ericsson and colleagues define deliberate practice as work that is well specified, feedback-rich, repeatable, and guided by a teacher or coach, and they argue that these features make skill acquisition more efficient than unguided repetition. Their framework was built from studies of expert performers and helped shape the modern debate over whether expertise comes mainly from practice quality rather than raw hours.

Meanwhile, a large meta-analysis from 2014 by Brooke Macnamara, David Hambrick, and Frederick Oswald titled ‘Deliberate Practice and Performance in Music, Games, Sports, Education, and Professions: A Meta-Analysis’ found that practice alone only explains part of the picture across different fields, leaving room for other factors as well.

The practical lesson is simple for anyone learning a new skill, whether coding, a language, a sport, or an instrument: replace some passive repetition with a slower, more deliberate loop of “try, get specific feedback, fix the mistake, try again.” It’s less immediately rewarding than clocking in the hours. But according to this data, this is what moves the needle.