Nylon Notebook: Updating Metrics and Sundries
By Seth Partnow
Nov 18, 2014; Sacramento, CA, USA; New Orleans Pelicans forward Tyreke Evans (1) on an offensive foul against Sacramento Kings guard Darren Collison (7) during the third quarter at Sleep Train Arena. The New Orleans Pelicans defeated the Sacramento Kings 106-100. Mandatory Credit: Kelley L Cox-USA TODAY Sports
Just a quick post to update[1. As well as use this post as a placeholder for a few metrics while we sort out some backend issues to allow for full automation and daily updates.] some of the metrics I’ve been tracking. If you use any of these metrics, bookmark this post because the graphs will be updated several times per week.
As always, there are a few caveats to this one:
- The metric doesn’t (directly at least) track deterrence, though that should be captured to some degree in contest%.
- It is not at overall measure of defense. 40% of the defensive value of a big man, is my best estimate, backed by some quick and dirty research.
- Contest% appears to be heavily scheme dependant. Anthony Davis spends time away from the rim, so Anthony Davis doesn’t contest as many shots at the rim as one might expect.
- Some non-bigs appear on the list, and a few bigs (Kenny Faried, come on down) don’t appear because they aren’t contesting enough shots at the rim to meet the minimum threshold of 5 contests per 36 minutes. For bigs meeting the minutes criteria, this is bad.[2. One of the things we’re working on is automatically separating players into rough positional categories to allow for a look at which wings and even guards provide a little extra value by defending at the rim at little. KJ McDaniels might not look impressive relative to big men, but from the wing, that kind of rim defense has value even if it’s wasted in the abyss that is Philly.] The average NBA big contests between 8 and 8.5 shots per 36, under 5 is quite hard to accomplish.
Nothing special to report here, other than Kobe keeps shooting – as friend of the blog Johannes Becker visualized the other day, in terms of the combination of time he spends on the ball and propensity to shoot, Bryant is an extreme outlier:
Just a note saying this data is super noisy this early in the season with the percentages of teams’ shots coming via assist chances bouncing several points up or down after each game. Some of Cleveland’s offensive struggle are indicated here – the Cavs are 4th in eFG% on assisted shots, but only 23rd in percentage of shots potentially assisted, which probably ties into their mediocre shooting so far. Worth remembering that while shooting is the most important of the “four factors,” these splits don’t take into account offensive rebounding, drawing fouls and avoiding turnovers.
Random Findings and Questions
“Reboundable misses” – much like having a low propensity for turnovers can provide a player with hidden value over and above what is indicated by some common efficiency metrics, a player misses shots in a way more likely to result in offensive rebounds, that’s probably a nice bonus. Huge caveat that this is very preliminary, with no indication of how much of the variation is a result of playing with teammates who are good offensive rebounders or for a coach who encourages crashing the boards, but there are still some interesting findings: Among players whose misses have resulted in less than 20% offensive rebounds, 9 of 21 are broadly describable as jump shooting bigs, perhaps lending credence to the theory that a “stretch” big is a tradeoff between floor space and rebounding.
Tyreke Evans (who was second on this metric in 2013/14) leads a host of “slashing” wings near the top. Bryant is also up there. The Kobe assist lives. This is certainly an area worth looking into in terms of if there is any “type” of offensive play or player that lends itself to higher or lower than expected offensive rebounding chances. To be examined further.
Shot Location and Defense – Just confirming some notions we’ve had about the desirability of certain shots. Even a wide open midrange shot hasn’t been especially great so far this year, relative to even closely guarded three pointers or contested shots at the rim:
A full examination of the data based on shot locations and contest distances yields some interesting points for discussion, but probably nothing definitive yet, as an example, the Knicks have been reasonably good at being near three point shooters (within 4 feet), but opponents are hitting 42.9% of these shots compared to a league average of 28.2% in the early going:
Whether this represents bad luck, the Knicks being particularly bad at closing out shooters or a mixture of both remains to be seen.
