Link to paper: Quantifying Shot Quality in the NBA Really cool paper based on SportsVU player tracking that proposes the metric ESQ (Effective Shot Quality) which quantifies the quality of a shot based on court location and proximity to nearest defender. Also proposed is the metric EFG+ (Effective Field Goal % - Effective Shot Quality) that attempts to quantify how "good" a player or team is relative to the quality of the shots they take. The team rankings for each stat is a testament to what Morey and co. are doing with their offensive philosophy... I'd love to see the RGV Vipers' numbers for this stuff. See spoiler below: Spoiler Complete paper list can be found here: Sloan Sports Analytics Conference Research Papers
I think the paper got the right idea. Does it also account for who that nearest defender is? Dwight Howard contesting your shot is a little different than Isaiah Thomas doing it, no?
The shot quality metric is simply dependent on proximity distance. I think the difference between Dwight vs. Isaiah would be apparent in their personal defensive EFG+ metric: If you, for example, control the shot at 50% expected shot quality, and the exact same shooter takes 100 of those shots against each defender and makes 55% (EFG) of them when Isaiah contests vs. 45% (EFG) when Dwight contests, then you'd have: Dwight defensive EFG+ = -5.0 (good) Isaiah defensive EFG+ = +5.0 (bad) See Pacers team EFG+ stats in this table: Spoiler
What happens to that scatter plot if we include Current Salary as a driver, too. The world screams for a true Money-ball chart that includes $ as an intensity factor. I demand a cost abatement curve.
On the other hand that effect might be cancelled out a little by the fact that a very good defender (Dwight) is probably more likely to be in close proximity to challenge shots anyway. That's just a guess though.
Seems the Rockets are best at creating high quality shots but we just don't capitalize on them by missing easy dunk/layups and missing wide open 3 pointers. This metric EFG+ also helps the likes of Dirk, LMA, and Durant. Tall players with good turnaround fadeaway jumper or simply players that'll shoot over you. They take high number of low quality shots (which in this paper is nearby defender) but still make a high percentage. Essentially of the top teams in EFG+ Heat and Spurs create good shots and make a high percentage. Blazers, Mavs, Thunder, and Warriors take tougher shots and make a high percentage. Rockets create great shots but don't make as high a percentage
Great point I hadn't considered... my initial thought was "aw that skews the metric" but then I realized it's simply the reality -- those guys are better shooters because of the fact that they can shoot over the average defender.
I think a marginal analysis of EFG + at the per game level would be very interesting. At is base level EFG + is basically how much a team "out performs or under performs" the expected points given up/ scored based off of the shot selections. Obviously the way teams are constructed there should be differences in the ESQ and EFG over the course of a season.. and the stat itself really doesn't tell why. Its is just a way to measure it. It would be interesting to see if high EFG + values forecast regression in certain areas ( although the quality of players can render this argument moot) Again it would be really interesting to see these stats compiled on a per game basis as a way of measuring consistency.
How do you figure? Looks to me like we give up too many quality shots based on us not being in the second chart.
We give up quality shots, but we defend them well. My first thought is that we funnel (read: play terrible perimeter D) wing players into the paint, where they take "good shots"... except for the fact that there's Dwight there. So the metric thinks they're good shots based on league average, but the Dwight effect makes them worse than expected. The other thought I have is how many open threes we tend to give up and get lucky on. Seems like, in my honest opinion, we've had some pretty good luck this season with open shots being missed by other teams. Probably a silly thought given the sample size... I'd like to see a stat on that too!
Not sure if serious, but it's not the same Adam Wexler. Paper writer (sideburns, etc.): Rockets' Wexler (distinct lack of sideburns, etc.):