Creative types all go through hot streaks of superior production

Poker players are familiar with hot streaks—times where it just seems the cards consistently fall their way. But we also talk about hot streaks when it comes to athletic skill, for things like hitting in baseball and shooting in basketball. Now, a group of researchers would like to add a few more activities where hot streaks take place: science, art, and movie making.

The researchers analyzed the production of high-impact work in those fields, and they find that it tends to cluster in a single time period. And, for most people, getting hot once is all they can hope for—getting hot twice is not common, and multiple hot streaks are extremely rare.
Random or streaky?

The work was an attempt to make sense of two seemingly contradictory results from past studies. Some had suggested peak productivity in creative fields tended to occur in a cluster around the middle of people's careers. Other studies had indicated that people's best work seemed to occur randomly throughout their careers.

To get a better sense of what was going on, the researchers chose three fields where they felt they could obtain some measure of a creative individual's success. Artists were included because it's possible to determine what price their work fetched at auction. For film directors, the researchers used IMDB ratings. And, for scientists, they looked at citations to their papers in the first decade after the paper was published. In all, they included 3,480 artists, 6,233 film directors, and 20,040 scientists in their analysis.

To determine whether an individual's best work clustered together in time (as you'd expect if they were on a hot streak), the researchers identified the most significant work of each person's career. They then compared when that happened to when their second-most-significant work was produced. When the difference between the two was plotted, it produced a curve with a strong peak centered at zero. In other words, the best and second-best work of these individuals' lives tended to be produced in the same year. Similar graphs were produced when second-best and third-best works were compared and when the best and third-best were compared.

The curves were symmetrical, indicating the best work was equally likely to come before or after the second-best. And if the person's output and dates were shuffled randomly, the pattern seen here vanished, as one might expect.
Feeling hot

This pattern is suggestive of a hot streak, a relatively short period where a person produces their best work. To test for this in another way, the researchers calculated the median quality of all of a person's works and then determined whether each piece was better or worse than the median. They then plotted this comparison over time to see if there were periods where a person's work consistently exceeded their average output. "Real careers," they conclude, "are characterized by long streaks of relatively high-impact works clustered together in sequence."

The hot streak is real, and it's even relatively common. Eighty-two percent of film directors experienced one, and the numbers were even higher for artists and scientists. Even the researchers were surprised by just how consistently it showed up: "Given the myriad factors that can affect career impacts, and the obvious diversity of careers we studied, the level of universality and accuracy demonstrated by the simple hot-streak model was unexpected."

But, while it's common in a group of creative individuals, it's not common within any one individual's life. Two-thirds or more of people in each field only experienced a single hot streak during their entire career. Having more than two hot streaks is rare.

The length of the hot streak varied a bit among the fields. For both artists and film directors, they tended to last a bit over five years. Poor scientists burned out fast, with the average hot streak lasting 3.7 years. The researchers checked, and the hot periods weren't associated with periods of higher productivity; as far as they could tell, they occurred at random throughout a person's career.

In fact, the authors aren't sure why anyone would go on a hot streak at all. In something like film, it's easy to see how success could breed more success, as a hot director would be offered better projects and resources to complete them. The same obviously isn't true for artists, though, and the slow pace of scientific funding and publishing would mean that the typical 3.7-year-long hot streak would be ending right about when a researcher obtained any resources as a result of the high-quality work that started the hot streak. So it's entirely possible that different factors control the hot streaks in different fields.

If so, it would make the consistent appearance of hot streaks in these three fields even more remarkable.