Posted by Neil Paine on March 4, 2009
Now, this is the kind of post I like to see... Basketball Geek's Ryan Parker (whose blog has quickly filled the "brilliant research" void left when Eli Witus shut down CountTheBasket to go to the NBA) took player on- and off-court shot location data and ran an adjusted +/- style regression to determine each player's individual impact on where his team's shots are coming from. It's especially useful because you can take the results and estimate what the offensive and defensive shot distribution of any theoretical 5-man lineup would be. For instance:
"Lets imagine a world in which the [07-08] Rockets loaned the [07-08] Celtics Yao Ming in exchange for some time with Kendrick Perkins. What sort of shot distribution would this lineup of Rajon Rondo, Ray Allen, Paul Pierce, Kevin Garnett, and Yao Ming have?"
Very cool, right? And definitely one of the more inventive ways I've seen of applying the on-/off-court regression methodology.