Fig. 2.

Career length, activity and productivity distributions. a The probability P(L) that an actor or an actress has a career of length L, estimated by computing the frequency histogram of the number of years between the first and the last recorded entry on IMDb. P(1) measures the abundance of “one-hit wonders”, namely the actors or actresses with IMDB records in a single year. A zoom for L ∈ [2, 10] in the inset shows that careers extending between 2 and 10 years are proportionally more frequent in women than in men. b Activity distribution P(s) estimated by computing the frequency histogram of the number of working years within each career (s ≤ L). Curves for actors and actresses are very similar and both exhibit a clear exponential tail, implying a "scarcity of resources". c Log-log plot of the total productivity distributions P(n) for actors (black) and actresses (blue). Both curves decay as a power law P(n) ~ n^−γ, where γ ≈ 2, revealing a Zipf’s law for the total number of acting jobs