Whenever I come across the words chess and practice hours in a paper, I immediately think of the late Anders Ericsson. Chances are, many readers will think of Malcolm Gladwell instead. That is hardly surprising as Gladwell made the idea famous in his book ‘Outliers’. The problem is that he simplified Ericsson’s work beyond recognition into the incorrect 10,000-hour rule. Ericsson himself spent years trying to correct that misunderstanding, most notably in his final book, Peak. But I’m afraid it did not really help. The simplified version persists, as many people still believe that becoming an expert is simply a matter of putting in enough hours. Ten thousand should do the trick. It is a great story. It is just not what Ericsson stated in his work.

Ericsson’s real question was much more interesting. To understand this, imagine two people who devote roughly the same amount of time to learning something new. Why does one continue to improve while the other seems to level off? You can see this in music, in sport, and probably in every domain where expertise matters. And as a matter of fact, also of course in chess. According to Ericsson, the answer was not that the better performer simply had worked harder. The difference was how they both had practised.

He described this as deliberate practice, an idea I also discuss in The Ingredients for Great Teaching. The principle is fairly straightforward. Instead of just doing more, you need to focus on what you cannot yet do well. That usually means:

• working with a clear goal,
• receiving immediate feedback,
• analysing your mistakes,
• repeating the difficult parts, and
• constantly pushing yourself just beyond your current level.

I admit that none of this list sounds particularly enjoyable. In fact, deliberate practice can often be pretty frustrating. But you need to realise that it is all demanding by design. More importantly, it is something very different from simply accumulating experience. If you practice the same note or chord over and over again on guitar, you will get well on playing that one note, but you won’t be able to play a song.

Unfortunately, that distinction almost disappeared from the public conversation. The message became “practice makes perfect.” Ericsson’s actual point was closer to “the right kind of practice makes perfect.“

But… was he right? Over the past decade, his theory has come under increasing criticism. Several meta-analyses concluded that deliberate practice explains far less of expert performance than Ericsson originally claimed. Other researchers argued that intelligence, motivation, personality, and genetic differences also matter. Ericsson pushed back against many of those arguments, maintaining that innate differences played only a limited role, such as length in playing basketball. The debate gradually became another chapter in the familiar nature-versus-nurture discussion.

But perhaps the disagreement was not only about the conclusions. Perhaps it was also about what researchers were actually measuring. Many of the studies critical of Ericsson did not measure deliberate practice as he had defined it. Instead, they counted practice hours, years of experience, or the number of chess games played. Ericsson repeatedly argued that this was like measuring how many hours someone spends sitting at a piano without asking what they actually do during those hours. Simply counting time, he argued, tells you surprisingly little if you ignore the quality of the practice itself.

So why am I telling you all this? Because a new paper by Daniel Southwick and colleagues on more than 44,000 Chess.com users is actually relevant to a debate that many people thought had already been settled. Sorry for the long detour!

The researchers analysed the behaviour of more than 44,000 Chess.com players over a six-month period. Rather than asking players afterwards how much they had practised, they used objective platform data showing exactly what players did. That made it possible to distinguish between several different activities: playing games, reviewing completed games, following instructional lessons, and solving tactical puzzles.

They found that players spent more than 90% of their time simply playing games. Which could sound perfectly sensible to many, thinking that if you want to become better at chess, you play chess.

Except… the data also showed that this was not the fastest way to improve. Reviewing your own games and following instructional lessons yielded roughly 3.6 times as much improvement per hour invested as simply playing another game. In other words, less than 10% of players’ time was devoted to these more targeted learning activities, yet they accounted for a disproportionately large share of the improvement. And more importantly, those activities closely resemble Ericsson’s idea of deliberate practice

Does that mean Ericsson has finally been vindicated? No, sorry, not really.

The authors themselves are careful not to overstate their findings. They could not measure intelligence, motivation, or personality, and they explicitly acknowledge that these factors almost certainly influence improvement as well. Their conclusion, therefore, is not that talent is irrelevant.

Their argument is more subtle. If previous studies treated every hour of practice as equivalent, they may have overlooked precisely the distinction Ericsson cared most about. In other words, at least part of the debate may have been driven by differences in how researchers operationalised the concept of practice.

So where does that leave us? Not with a revival of the 10,000-hour rule. That remains a misleading summary of Ericsson’s work. What this study does remind us is that scientific debates often hinge on definitions. If researchers measure something different from what a theory originally intended to describe, it should not surprise us that they reach different conclusions. That may not be as catchy as “10,000 hours”, but it is probably a better description of how science actually moves forward.