When bowlers talk about having “hot hands,” they may be onto something.

A new study by Gur Yaari, a Yale pathology postdoctoral associate, and mathematics professor Gil David offers evidence supporting the popular notion of “hot hands,” the idea that an athlete’s performance improves during winning streaks. Published Jan. 12 in the journal PLoS ONE, the study found that professional bowlers had streaks that were more than just random variation. The authors said that this study contradicts previous research which generally found the idea of “hot hands” to be a mistake of perception.

For example, 1985 study that Yaari called “influential” determined that each attempt to make a free throw in basketball had a completely independent chance of success. This would imply that a hot streak is a variation on the gambler’s fallacy, the false idea that past results in independent trials, like flipping a coin, will change future results.

In their recent publication, Yaari and David claim to have found evidence for “hot hands” in bowling.

“It’s not so surprising that luck doesn’t perfectly model human performance in sports,” Yaari said.

The study used data from over 40,000 games played by the top 100 bowlers from the Professional Bowling Association.

They found that bowlers who bowl more strikes than their average in the first eight frames will be extra likely to bowl strikes in the last two frames. But Yaari said that the results do not necessarily imply a direct causal relationship between any two successive strikes. Or, in other words, bowling a strike on one frame will not directly increase the likelihood of bowling a strike on the very next frame.

Instead, bowlers appear to have identifiable “good games” and “bad games,” which are strong enough to suggest that each attempt is not completely independent, like successive coin tosses.

“There is a large amount of evidence that suggests that the ‘hot hand’ phenomenon that sports fans [think] they are observing nearly every time they watch a sporting event does not exist,” said Jonathan Koehler, a professor at Northwestern Law School and expert in behavioral sciences. He said he had not read this particular study, but cautioned against over-interpreting “hot hands” research because, on the whole, the phenomenon has proven remarkably difficult find scientifically, despite how often people tend to identify it in sports.

Sidney Redner, a physics professor at Boston University, said the study’s analysis looked legitimate, but added that the effect appears to be much subtler than people often realize.

Yaari agreed, and he said that the hot hands effect should not be overstated.

“People have in mind causality, but it may have more to do with streaks that don’t involve explicitly psychological feedback,” he said. The study refrained from making any conclusions on what might be causing the effect.

Yaari’s previous study, published in PLoS ONE last October, found a similar effect in basketball free throws, and he said that the findings might be relevant to any task, like playing video games, that requires intense concentration.