WP48 Misleading?

I wondered back in April how accurate it would be to assume a player's WP48 wouldn't change when they change teams. I thought it would probably be biased heavily upward for players who take a lot of shots because each game only has so many possessions.

You can break a players value into possessions used and efficiency per possession. Some players are generating value largely from the former, so when they transition from a team where they are the star to a team where they are second (or third) fiddle their value drops.

Look at what happened to the Heat:

	TS%	Shots	WP48
James	-0.031	-3.35	-0.117
Wade	0.011	-2.85	0.008
Bosh	-0.028	-3.8	-0.102

In the Heat's case part of the WP48 decline is a small decrease in efficiency. The bigger part of it is a decrease in the number of shots attempted by over 10% each. For James and Bosh though, I suspect, that most of the WP48 drop is accounted for by declines in rebounded, steals, etc.

This picture tells the story about the Big Three: