breaking news

Jazz Individual Shot Selection Stats

December 13th, 2013 | by Ben Dowsett
Photo by Garrett Ellwood/NBAE via Getty Images

Photo by Garrett Ellwood/NBAE via Getty Images

A few weeks ago, with the Jazz still in the hunt for the title of WOAT (Worst Offense of All Time), we took a look at a fundamental question: were Utah’s plethora of offensive issues the cause of poor systems, poor execution, or some combination of the two?  With help from, we took a look at how the Jazz compared league-wide in terms of points per shot attempt, both expected and actual.  My previous piece with a more detailed explanation can be found here, but a quick recap:

Using shot location data from as early as the 2000-01 season, Hickory-High was able to calculate an average point-per-shot attempt value for five different shot distances.  Using these averages combined with teams’ and individuals’ locational breakdowns, they can then generate an expected total points-per-shot value for any team or player by simply assuming league-average percentages from each distance.  They can then compare this with the actual points-per-shot generated by these teams and players – a simple and useful way to figure out which guys get the best shots, and which guys are simply better at making the shots they do get.  For this piece, some of the data was actually generated by myself and the wizard Joe Kolassa using Hickory-High’s formulas (kindly shared with me by founder Ian Levy; after you read this, go follow him and Hickory-High on Twitter – it’s well worth your time).

Unfortunately, the results for the Jazz were less than glowing in my initial write up.  While they were nearly league-average in XPPS, their actual PPS was dead last in the NBA and one of the worst marks in league history since shot location data began being tracked.  The translation is simple: Utah was at least moderately capable of generating high-value shots, but their execution was miserable.

But while they’ve not morphed into the Heat or Thunder overnight, the Jazz have started to show some real improvements lately – at least according to more common measures like points-per-100 possessions, where (as I noted last week) the Jazz have approached league average since Trey Burke has been in the lineup.  Ty Corbin’s move to smaller units, particularly those featuring Marvin Williams, has also had a noticeable positive effect on the offensive spacing.

And with that in mind, I’d like to re-visit the theme of XPPS, this time with an emphasis on the individual Jazz players themselves.  But first, a quick look at how the team as a whole has progressed in the weeks since I first wrote about this subject:

Interestingly enough, Utah’s XPPS has dropped noticeably compared to the rest of the league during this time.  While the Jazz during their awful stretch were ahead of certain elite offenses (but, of course, well behind in actual PPS), this is no longer the case – they’re down to 25th in the NBA.  Execution-wise though, as one might have expected, the Jazz have improved.  It’s a good thing, too, because they couldn’t have been much worse.  They’re only up to 26th league-wide, but it’s a far cry from the mark that had them competing with Charlotte for one of the worst ever under a month ago.  This was to be expected; the Jazz aren’t that bad, and their actual and expected numbers inching closer together signals more stability.

Just before we get into the individual members of the roster, do note that these numbers (for now) aren’t manipulated for time periods, and will always reflect the full season.  And because the Jazz were somewhat in chaos for several of the first few weeks due to injuries, inexperience and one of the largest offseason rotational turnovers in the league, their data will be weighed down a little – not enough to disrupt any larger trends, but still something worth noting. The numbers:

Jeremy Evans 1.034 1.218 0.183
Brandon Rush 1.020 1.190 0.171
Marvin Williams 1.009 1.147 0.138
Michael Harris 1.011 1.108 0.096
Enes Kanter 0.995 1.045 0.050
Derrick Favors 1.079 1.114 0.035
Richard Jefferson 1.059 1.069 0.010
Utah Jazz 1.030 1.022 -0.008
Alec Burks 1.043 1.022 -0.021
Diante Garrett 0.939 0.911 -0.028
Trey Burke 0.991 0.959 -0.032
Gordon Hayward 1.018 0.986 -0.032
John Lucas III 0.986 0.836 -0.150
Ian Clark 1.021 0.805 -0.216
Rudy Gobert 1.272 0.854 -0.419
Jamaal Tinsley 1.014 0.450 -0.564
Andris Biedrins 1.421 0.824 -0.597

And as the team numbers seemed to suggest, things have indeed begun to even out somewhat.  Where nearly every Jazz player carried a negative differential between expected and actual PPS a few weeks ago, they’re now evenly split – seven players with positive differentials, seven with negative ones.  For such a young team, this is actually a relative positive; at least half their players are capable of making shots at above their expected rate.  Jeremy Evans is the largest standout in this area, although his +.183 differential is nowhere close to the league-leading rate of over .400 he sported for several weeks before falling back to earth recently.  Sample size and more minutes against starters caught up to him, but the numbers are still a very encouraging sign of a breakout year for Evans – generating shots at nearly league average while proving himself a much more effective shot-maker than many had given him credit for.

Other positives include Derrick Favors and Richard Jefferson, both of whom are producing expected and actual numbers over league average despite all the struggles Utah has had.  Favors, in particular, has been choosy about his shots – over half his total field goal attempts come from within the restricted area, per, exactly what Jazz fans would hope to see as he continues to, ahem, perfect his jumper.

The biggest plus for Utah, though, has been Marvin Williams.  Thriving in his new role as a stretch-four in Corbin’s new small-ball units, Williams has opened up all sorts of new possibilities.  The raw numbers place him third on the Jazz for expected/actual differential, but the two players above him (Evans and Brandon Rush) are both on a much smaller sample size and shoot far less while on the court.  Well over half his field-goal attempts come from beyond the three-point arc, and those watching every game will know how high a percentage of these looks have ranged from “barely contested” to “wide open” given the excellent spacing generated by these smaller lineups.  One interesting note: Marvin is producing such above-average PPS numbers despite a surprising lack of proficiency on the corner three – he’s shooting just 18.2% from the corners on just over one attempt per game.  Compare that with 47.7% on all other threes, on nearly triple the nightly attempts…if Utah could work to move Marvin to the corners more frequently and shake his rust off from there, they could improve his efficiency even more given the increased value of corner threes.

Of course, there are some negatives as well, though nowhere near the calamity of mid-November.  Trey Burke and Gordon Hayward are both slightly below-average for both expected and actual PPS, perhaps a slight surprise for Burke, at least.  Mid-range jumpers are the main culprit for these two, as both are attempting more mid-range than any other type of shot so far this year despite neither shooting better than 37.0% from there.  Hayward has been particularly bad at times with his shot selection – the pull-up 20-footers fading away over a screen with 15 seconds left on the shot clock need to stop, and I expect Hayward himself knows this.

But of course, what analytically-driven piece would be complete without a cameo appearance from the king of advanced stats comedy, Andris Biedrins.  In his 42 glorious minutes on the court this season, Biedrins has attempted a single field goal, which he made.  So with just one shot attempt, and two points generated from said shot attempt, he should top the actual PPS list, right?  Of course not, he’s Andris Biedrins!  Because Hickory-High’s formulas also account for free-throw attempts (a detailed description below the tables here), Biedrins’ 1-for-6 mark at the line this year (at least he made one!) drops his actual PPS to a dismal .824, with a negative expected/actual gap of -.597 – one of the 15 worst marks of roughly 450 NBA players.  It’s just not fun unless there’s something to laugh at Andris Biedrins for; thanks for never disappointing, buddy.

As I noted in my original piece, this type of data can only paint a certain portion of the larger picture.  Sample size must be accounted for, as guys with a small number of shot attempts can have their data skewed by a few lucky makes or misses.  But when viewed in the proper context, XPPS is a valuable tool for examining problems in system versus problems in execution.  The modern NBA game is trending toward these sort of metrics, and it’s good to see the Jazz making some strides.  I’m intrigued to see, as the year goes on, how they’ll continue to improve with Corbin’s newfound offensive creativity.

Leave a Reply

Your email address will not be published. Required fields are marked *