I'm doing some work on ref ratings, and have gotten to the point that I am trying to filter out crew effects. In other words, I'm trying to control for the situation in which a ref "earns" a bias by working with another ref who really does all the work of calling for the bias.
Before I start seeing how often crews are intact, I thought I'd just ask if anyone knows or has broken this down. How often do refs rotate from crew to crew? If a ref refs 65 games a year, how many crews has he typically been a part of?
Clear biases in zebra behavior do exist, and the only NBA approach I've experimented with is 2-against-1, where two refs have the same strong tendency and the other does not.
This season starts out with backup refs. But in any case, zebra tendencies are not passed on from one season to the next. I'm not sure why. Maybe they're short term fluctuations, or maybe the extreme refs become aware of their tendency and adjust for it.
I know the refs are graded on their calls, and if they lean too far one way they get feedback about it. I'm testing a theory that some refs favor teams that prefer fast or slow paced teams. These are smaller samples, but I hope a more stable bias.