advanced analytics do work for sports that have finite outcomes and are open to little subjectivity (baseball, football) in terms of measurement. The art, for those sports, comes in the interpretation and coloring of the story that you think the numbers tell.

In basketball, there is so much situational ambiguity that I am skeptical you can build even a basic data set that tells a remotely concrete story about a player, coach or team's effectiveness or execution. It's too free-flowing, too interchangeable, too random. When continuation fouls, for example, entered the game, everything went out the window.

That said, it's not all down to mystical ability. Hollins' Kobe clutch argument, for example, has been explained away as a simple numbers probability; he makes the shot because he's given the most chances. And he's given the most chances because he makes shots, or we believe he does at a higher rate than an alternative shooter.