Correlation what? I've got your linear relationship right here.
Tom Ziller does his usual fantastic job with fancy statistics and uses the terms "correlation coefficient' and "linear relationship" in a non-ironic way without sounding absurd. The main point is that 3 pointers are more important than ever. I suppose that it is a good thing that the Celtics went out and got two of the better shooters in the game (despite losing one of the all time best).
However, one conclusion that I wanted to highlight is the much maligned "long two."
Three-pointers rule the land. It's also worth [noting] the biggest problem with long-two pointers: that they are not three-pointers. The share of FGAs taken as long two-pointers has a -0.44 relationship with actual eFG. Shot shares at the two other inefficient ranges -- short and mid -- also have negative relationships with actual eFG, but with much, much smaller correlation coefficients. Why are long two-pointers such a problem? Check out the correlation between rate of long twos and rate of threes: -0.57. In other words, very few teams take lots of long twos and lots of three-pointers. So every long two is basically a three-pointer not taken. And three-pointers are important.
I know that most of you don't know what Mike Zarren looks like, but the mental picture of him trying to explain this to Kevin Garnett has me in stitches at the moment. That's not to say that KG isn't a student of the game. I just don't think you are going to convince that old dog to try too many new tricks. Of course, if he took an extra step back on a few of those long 2's, it wouldn't bother me one bit. I'm just not signing up to be the one to tell him.