Page 1 of 1

Turning Four-Factors Into Efficiencies (bbstats. 2010)

Posted: Fri Apr 15, 2011 9:27 am
by Crow
Author Message
bbstats



Joined: 25 Apr 2010
Posts: 46


PostPosted: Thu Dec 23, 2010 6:59 pm Post subject: Turning Four-Factors Into Efficiencies Reply with quote
We can do this inductively, as Ed Kupfer and others have toyed with, by fitting linear regression models with intercepts of zero.

Or, you can try it deductively, as I have attempted to do in the past (where I include FT% and FG%).

My current method uses .475 for the *FTR term on the bottom, and I use a 1.07 coefficient on the OR%*(1-FG%) term on the bottom as well.

If we assume the team's estimate for possessions to be accurate, this gives an R^2 of 1. Including opponent's estimate for possessions to get an average, the R^2 value dips a bit (I haven't quantified this yet, sorry).

I just thought I would point out that the four factors basically tell us precisely how many Field Goals are attempted per possession:

Code:
FGA/poss = (1-TO%)/(1+.475*FTR-1.07*OR%*(1-FG%))

_________________
http://thebasketballdistribution.blogspot.com

http://twitter.com/bbstats