A machine-learning method that can predict whether an actor's most productive year of work in film and TV has already occurred or lies ahead is presented in Nature Communications. The study reports that the most productive year tends to be towards the beginning in an actor's career, and this effect is more pronounced in the case of female actors, who are also more likely to have shorter careers than male actors.
With a 90% unemployment rate and only about 2% of screen actors making a living out of acting, simply having enough work (sustained productivity) can constitute success for most actors.
Lucas Lacasa and colleagues used a worldwide database to study the temporal patterns of productivity over the careers of more than two million screen actors between 1888 and 2016. The authors find that most actors have very few credited jobs during their careers, although a few actors have over a hundred, indicative of a 'rich-get-richer' mechanism with respect to the allocation of work. They also report that although the percentage of an acting career spent working is unpredictable, active years are clustered into hot and cold streaks, where actors are more likely to work in a given year if they worked the year before and less likely if they did not. The authors show that it is possible to predict with an accuracy of 85% whether an actor's most productive year of work has already passed or is still ahead of them, based on their work history.