More SportVU Aided Observations on Total Offensive Load
Initial discussion of "True Usage" metric here
I've been noodling around some more with the data from my look at total offensive load (still would LOVE a better name for the stat, suggestions please. EDIT: Colin suggests "True Usage" done and done), and seeing if there are some interesting tidbits to be gleaned. I mentioned a few of them in that post, but here are a few more. Further, I cleaned up the numbers a little (stupid spreadsheet errors!). All data is through Monday's (1/6/14) games.
First things is that I have averages for the overall sample and by each rough position category:
| Player | OffLoad | To/Load | PlayEFG |
|---|---|---|---|
| Overall | 34.37% | 8.80% | 0.529 |
| PG's | 46.39% | 8.28% | 0.555 |
| Wings | 32.17% | 8.72% | 0.525 |
| Bigs | 29.93% | 9.41% | 0.510 |
Some notes here:
- The average traditional usage rate for each positional grouping is right around 20% - just over for PGs and Bigs, just under for wings. Extrapolating from those averages, it's easy to see that generating "assist opportunities" is a positive for an offense - the higher load for PGs than wings and for wings than bigs comes from assist chances rather than additional shots, with a commensurate increase in PPP.
- The turnover rates make intuitive sense - as I mentioned in the initial post, point guards are point guards because they handler the ball well, meaning lower turnover rates despite the greatly expanded role in the offense. In raw totals, point guards have always had high turnover amounts, but I think this data shows that this is a result of the sheer amount of time they have with the ball and the responsibilities they have in terms of making offense happen. I wish I had SportVU data from Steve Nash's MVP seasons because even though he was in among the league leaders in total turnovers, I'll wager that he was probably one of the best players in the league in terms of per possession "used".
- This jibeswith a look at leaguewide Synergy data, which showed plays "created" largely without the ball (spot ups, PnR rolls, cuts) to be more efficient than possessions used by the player with the ball initially (isos, PnR ball handler, post ups).
- This makes me wish NBA.com had publicly available team totals for the SportVU data, as I'd like to see the overall correlation between assist chances and ORTG or EFG%, as well as see if there is any systemic relationship between number of passes and assist chances, drives and assist chances and so on. This actually has me mildly giddy, because one of the things that has long been on my wish list is the ability to model "shot creation" in some way, but I digress.
In less shall we say esoteric terms, the data also lets us look and see which guys are the biggest "black holes" in the league where if the ball gets to them, don't expect it to come back out. It also lets us look at who the most "pass first" guys (excluding PGs who would naturally dominate the list) in the league are.. Note the Asst Chances/Shot column does not include turnovers in the possessions used whereas TOs are included in the Asst Chances/Load calculation)
Black Holes
| Pass First
|
- Most surprising name on the first list is Chris Bosh, as every one else is primarily a finisher (even low usage players like Singler, KCP and Ross). As he has always been regarded as a good passer, I think this just indicates how little he has the ball in his hands in Miami.
- While Anthony Davis's spot is understandable given how and where he tends to get the ball, but his spot alongside Ryan Anderson's high ranking continue to make me skeptical about New Orleans's offensive scheme. Combined with Monty's seeming unwillingness to actually play his best players and the Pellies' pitifully low propensity to shoot from 3 (second lowest rate of 3's attempted in the league ahead of only notoriously jumper-challenged Memphis) despite multiple willing and able gunners, Williams might be the most actively harmful coach in the league. He's by all accounts a very nice and decent guy, and might even keep his job given the obvious and fair excuses given by the 'Cans rash of injuries to their core players, but I'd really like to see that team in the hands of someone more capable.
- A few of the more surprising names on the second list (Bayless, Lou Williams , Monta) are explainable by the amount of time each spends in a point guard-like role, as do LeBron, Gordon Hayward and Lance Stephenson. Just missing the second list (and easily top 20 if turnovers are included): Kyle Korver. The man is a golden god.
- On the flip side, Ricky Rubio looks to shoot the least of any player in the league as the only player with more than twice as many assist chances as shooting possessions. A-hem. Look at the bucket, Rick.
- On the other hand, Kyrie Irving has the lowest shot/assist chance ratio of any regular PG in the league.
Some interesting eyeball results looking at the data for teams or sets of teammates:
- Charlotte has two of the four "chuckingest" PGs in the league in Kemba Walker and Ramon Sessions.
- Speaking of Kyrie, among Cleveland's regulars, only Anderson Varejao has a significantly higher pass/shot ratio than his positional average (Jarrett Jack is slightly above average for PGs, but only slightly). With more talent on the wing after the addition of Luol Deng, it will be interesting to see if their ball movement improves.
- Speaking of Deng, don't be surprised if Chicago's offense falls off even further - he was using a large load and doing so extremely effectively (caveat about my doubts over the accuracy of "PlayEFG") - how much of the slack will Jimmy Butler be able to pick up?:
| Player | OffLoad | To/Load | PlayEFG |
|---|---|---|---|
| Carlos Boozer (CHI) | 34.83% | 11.55% | 0.464 |
| Taj Gibson (CHI) | 30.12% | 12.45% | 0.426 |
| Joakim Noah (CHI) | 31.00% | 9.71% | 0.531 |
| Jimmy Butler (CHI) | 28.10% | 8.46% | 0.714 |
| Luol Deng (CHI) | 39.20% | 9.10% | 0.728 |
| Mike Dunleavy (CHI) | 29.37% | 7.57% | 0.499 |
| Kirk Hinrich (CHI) | 38.73% | 8.22% | 0.513 |
- Steve Blake put up some silly numbers before his injury. Interested to see if Kendall Marshall (a player I do not think especially highly of on the NBA level) approaches Blake's passing effectiveness once he gets a big enough sample size.
Quick get well wishes to Rob. We were going to podcast today, but he's well under the weather. Having recovered from what my father-in-law speculates was Swine Flu last weekend, wish him a speedy recovery.
RSS Feed
