The San Antonio Spurs’ Danny Green broke the record for three-pointers in the NBA Finals last June, making 25 of his 38 attempts in the first five games against the Miami Heat. If you hadn’t heard of Green before, you were not alone—he’s a solid player, but no all-star. The record-setting five shots per game was more than twice his regular season average. To essentially all basketball fans and analysts, it was obvious that Green was “hot”—performing at a much higher level than his norm.
And yet, Economics Nobel Laureate Daniel Kahneman says “the `hot hand’ is just a figment of the imagination.” This claim has become part of the conventional wisdom in behavioral economics. It is often used, by academics and talking heads alike, to provide a stark illustration of cognitive bias. New York Times columnist David Brooks wrote last year: “When a player has hit six shots in a row, we imagine that he has tapped into some elevated performance groove. In fact, it’s just random statistical noise, like having a coin flip come up tails repeatedly.” Former Treasury Secretary Larry Summers, who almost became the Fed Chairman, was quoted (also in a 2013 New York Times piece) saying, “The answer [to the question of whether players sometimes get hot] is no.”
The idea that there is no hot hand is based on a body of research, beginning with Amos Tversky (Kahneman’s long-time collaborator) and co-authors’ work in 1985, that failed to find evidence of streaky shooting in a wide range of settings: field goal shots taken during games, free throws in games, and shots taken outside of games. The claims are consistent with an even larger body of psychology research indicating that people, in general, often mistakenly see patterns in random data. Researchers came to a consensus that basketball “hotness” does not exist.
If success begets success in basketball, then why not also in other situations? This point, while speculative, could be relevant to a variety of policy questions.
This conclusion was largely ignored by fans and players. Red Auerbach, the great Boston Celtics coach, is reported to have said in response to the original research: “So he (Tversky) makes a study. Who cares?” The conclusion was at odds with decades of established basketball strategy. Bill Russell, the Celtics star who, under Auerbach, won 11 championships in his 13 seasons, said that one of the secrets to his teams’ success was identifying which player was hot and feeding him the ball.
Academics may have assumed this refusal to accept the research was due to a combination of ignorance and stubbornness. But recent work, including three peer-reviewed papers based on our own research and a new paper being presented at the high-profile MIT Sloan Sports Analytics Conference this Friday, has shown that the laymen seem to have been right all along.
New data, including both larger sample sizes and new variables (e.g. shot difficulty and defense intensity), combined with new insights into the interpretation of the statistical analysis, indicate that, yes, there is a hot hand—and it may, at times, be sizable. The absence of evidence in the previous research turns out to have been quite weak evidence of absence.
That said, the “hot hand bias”—the tendency to impulsively infer a player is hot, based on limited data—is still alive and well. The behavioral researchers were correct to identify this as an important cognitive error. But this does not mean there is no hot hand at all. A player who hits a few tough shots in a row may indeed be the best option for the team’s next shot.
Aside from enhancing our understanding of basketball, why is this new hot hand research important? It indicates the previous work was an interesting case study of scholarly overreach. Saying “there is no hot hand”—that virtually all players and fans were wrong—was much more attention-grabbing, and thus, perhaps, appealing, than simply saying there is a more subtle hot hand bias.
It is poignant that behavioral economics and psychology researchers seem guilty of the overreach here because a) they should be especially aware of the bias to exaggerate and believe what we want, rather than what is supported by the data, and b) those researchers should have been relatively confident in the existence of the hot hand, and thus skeptical of the initial research interpretation, since becoming “hot” is likely largely a psychological phenomenon.
This relates to a second, deeper implication of the new research: the potential importance of psychological factors, confidence, and momentum in performance in a range of contexts. If success begets success in basketball, then why not also in other situations? This point, while speculative, could be relevant to a variety of policy questions; for example, at what age are returns to schooling highest? Momentum effects (more likely to exist in general, given that they exist in basketball) would support focusing public and private investment in schooling at earlier ages, consistent with the recent research by another Economics Nobel Laureate, James Heckman. Better results early can give children confidence, making them more likely to achieve better results later.
While the recent labor economics literature has recognized the importance of negative psychological effects (anxiety and choking) on performance outside of sports, the positive effects are the other side of the coin, and should also be studied.
Last, it is worth explicitly noting the delicious irony of this story, and the implications. The claim by Ph.D. researchers and pop-intellectual authors that all fans and players and coaches were off-base was in fact itself off-base. This lesson supports the idea of there being great wisdom in the collection of human experience. Academics and others should not forget to have more faith in this collective wisdom, and to be more skeptical of new, counter-intuitive claims.