Love it or hate it, I did some work to come up with a variable equation for expected corsi using the variables we've been discussing in this thread. I have just over 1600 data points using defenseman since the 08-09 season who have logged over 250 minutes total TOI.
My r-square was low at .38; however, my p-values were all wildly close to 0.
Much to my surprise, Girardi's eCF% comes out to 47.13% while he's currently at 46.41%, a discrepancy of .72%.
Certainly underperforming, no question; but to attack his CF% and call him a possession anchor when his expected CF% is just 0.72% higher than his actual? Well...
For the rest of the Rangers D...
For curiosity sake, Barret Jackman leads the way in differential, destroying his eCF% of 44.91 with a real CF% of 53.1. Flipside, Andrej Mezsaros is terrible, with an eCF% of 47.2 and an actual CF% of 37.95%. The worst non-Sabre of the bunch is Nate Guenin. eCF% of 47.59% and an actual CF% of 40.64%.
eCF% has a range of 41.88% (Gorges) to 57.88% (Kindl).
Just realized there are two more variables I could've used. I'll do that later or tomorrow.
Thanks to the Rangers not playing well enough to fully grasp my attention today, and wanting to put the game on in the background, I've gone ahead and added those variables...
I'm much more satisfied with this run than the previous one.