Our research constitutes an advancement and attempts to answer an important question on basketball metrics: how to measure regularity in terms of points scored and which factors may in
uence this regularity. This question has been understudied in the specialised literature despite its importance. We employed the median absolute deviation (MAD) as a robust measure of regularity. After the analysis of the performance of 27 NBA players who played 82 games in the 2007 regular season, we have demonstrated that some players are much more regular in terms of point scoring than others, and when players are better scorers they become more irregular scorers, i.e. their performance is less predictable. In addition, some players at the half-level of regularity are "outliers" in the sense that some peaks in their performance may in
uence the variance of their respective distributions. Finally, the pattern of points scored per minute in a short series of games is very dicult to predict without taking into consideration additional information regarding in
uential variables; there is little evidence that NBA players are subject to a momentum e ect.