Has anyone done a data-based study of coaching behaviors? For such things as:
speed to calling a time out after falling behind 6-10 points and shot selection in the few minutes after that.
degree of rotation time alteration when a guy has a streak of missing 3-4 more shots than makes
degree to which average sum of player rebounding on the court is increased if actual game rebounding is unfavorable by some amount
etc.
Seems like many things would be worth checking if coding speed and talent is high.