Has anyone done this yet? Calculated adjusted +/- for "clutch" possessions? I did a quick Google search for "clutch +/-" and didn't turn up any hits, so I'm thinking maybe not. I think you should probably also take this study with a big grain of salt. Ah, who am I kidding. You know you're going to use this to point out that Kobe Bryant only ranks #40 and Monta Ellis only ranks #86, and that Dirk Nowitzki ranks #3 and LeBron James ranks #10. You know you're gonna do that, so I won't try too hard to stop you.
One of the problems with "clutch" is simply defining it. I'm using the 82games definition:
4th quarter or overtime, less than 5 minutes left, neither team ahead by more than 5 points
I'm using the past "2.5" seasons of data. In that time, there are 136 players who have more than 250 "clutch" possessions, which I'm showing here (obviously, that threshold is also arbitrary). Like I said, make of this what you will. (Oh, and for the statheads, I did this using cv.glmnet with and , which was found to give the minimum C.V. error). I'll point out that there are quite a few Cavs up at the top, which I'm guessing is mostly due to LeBron, and that the regression is really splitting the credit there. I'll do a follow-up post visualizing clutch units, because my hypothesis is that there is going to be a lot of collinearity (i.e. very sparse substitution patterns in clutch time).
[googleapps domain="docs" dir="spreadsheet/pub" query="key=0Al6a2ecvJfTidGtRTHhTdXI5WURRT3I3eXpYWkJnU2c&output=html&widget=true" width="380" height="525" /]