One thing I should have done in my study was take teammates more into account.  I took a look at the top subs of some of the players, which was hinting at the issue, but I didn’t delve into it further.  A player’s impact on his teammates’ three-pointers could be deceiving if he tends to be on the floor with certain teammates frequently.  For example, I pointed out that one of Dwyane Wade’s top subs was Daequan Cook, a long-range bomber.  This makes Wade’s impact on team three-point shooting look smaller than it actually is.  What about the players who often play next to long-range shooters?  Dwight Howard plays a lot of minutes with Rashard Lewis, so it makes sense that the Magic shoot a lot of threes when Howard is on the court (but not necessarily because of Howard).

To get around this, you have to look at it on a player-by-player basis.  Below I have tables for four of the players in the original study.  Each table includes the superstar’s impact on the three-point attempt frequency and efficiency for every player on his team.  At the bottom, I calculate a weighted average.

When we use this approach, the results tell us something different.  Dwight Howard still has a large impact, but it’s no longer the biggest of the bunch.  On average, Dwyane Wade increased each Heat player’s three-point attempt percentage by 11 percent.  Even Steve Nash, who I originally said had no effect, creates an average increase of 4.1% for his teammates.  Additionally, Nash makes his teammates shoot three-pointers more efficiently, raising their 3PT% by 4.3%.

This new data may change the specifics, but it further confirms the original idea: superstars create good three-point looks for their teammates.