The NBA may take the summer off, particularly the typically dead month of August, but nerdiness has no “season”, so neither do I. Last week I used this space to look at some contextualized elements of last season for the Jazz through the lens of SportVU tracking data, namely their shooting prowess (or lack thereof), passing and elements of rebounding. The week before, I used shot charts created by my Nylon Calculus colleague Austin Clemens to accentuate individual shooting and shot selection among vital Jazz pieces going forward.
In keeping with the theme of breaking down last season with a keen eye toward next, today I’ll be highlighting another remarkable project pioneered by a Nylon Calculus writer, Matt D’Anna. The data is affectionately referred to as TeamSPACE – an extrapolation of Austin’s excellent charts focused not on individuals, but rather on entire five-man units during their periods on the floor together over a given season. Before I say anything further, check out an example in the form of last year’s New York Knicks:
Matt used the Knicks as part of his inaugural TeamSPACE introduction, where you can also find more detail on his methodology. In a nutshell, though, it’s easy enough to decipher; each player within a given lineup is marked with a color in the lower right corner, and said colors correspond to their shot clusters on the court. Again, these are shots taken only while this particular five-man unit shares the floor. One extra nugget is the presence of weighted values for made shots over missed ones – clusters are primarily based on volume of attempts from the given areas, but Matt included slightly heavier weights for makes as compared to misses, so a player who chucks away repeatedly from an area they never connect from will show a lighter cluster there, or even in extreme cases perhaps no cluster at all.
Today will be Part 1 of my investigation of Jazz shooting within lineups last season, where I’ll break down data from their most frequently-used lineup of Burke-Hayward-Jefferson-Williams-Favors[ref]Check back next week for analysis of the “Core Five” (Burke-Burks-Hayward-Favors-Kanter, for those under rocks recently) lineup over the course of last season.[/ref]. Let’s take a look at the chart:
Part of what makes Matt’s work so interesting is that we can gauge both team and player context from the same visualization without sacrificing quality in either. Utah’s chart from last year is a prime example – within this lineup, Marvin Williams obviously stands out. Outside the Restricted Area (where all teams naturally clump up), he has the largest and most concentrated clusters of shots, particularly from beyond the 3-point arc. He has very few random blotches outside his preferred shooting areas, save for a few clusters from midrange that were frequently one-dribble step-ins after a close-out to the 3-point line from a defender. This type of clumping is almost always desirable, an indication that a player has identified his strongest areas and is taking steps to get to them regularly within the offense (more on this in a little).
Richard Jefferson was another positive through this lens, even more selective than Williams and rarely shooting from anywhere but beyond the arc or at the rim with this lineup. Derrick Favors also showed glimpses of range from the right block baseline up to free-throw line extended, but spread-out clusters here indicate that he hasn’t quite found a “sweet spot” or two to lean on, something he’ll need to work on if he wants to continue expanding his game away from the basket. And while I wish I could say more positive things about the guards’ showing here, the only major plus for Burke and Hayward was that the latter at least kept his chucking to some of the same general areas.
This shouldn’t surprise anyone by now, so I’ll just say it: This form of shot chart is just as ugly as any other metric we’ve attempted to analyze Jazz shooting with. I talked a couple paragraphs above about concentrated clustering being almost universally positive within this context[ref]To me, exceptions here would be players like Steph Curry, so deadly from anywhere within 30 feet that a spread-out chart is no problem as long as a majority of the shots are at least somewhat open.[/ref], and Utah’s visualization displays nearly the exact opposite. Compare their chart above, for example, to last year’s NBA champion Spurs:
Obviously, comparing Utah’s performance to a legendary offense is foolish in a vacuum, but in this case it helps partially illustrate some of the issues they had shooting the ball and as an overall offense. Look at how much more spread out the Jazz were in their shot selection, and just how beautifully concentrated San Antonio’s individual clusters were within each player’s preferred zone. Kawhi Leonard, he of multi-talented Finals MVP-winning pedigree, was the only starter with even slightly varied clusters, and you might say he had an OK year last year – no one is complaining about his variety. And outside Leonard, it’s the embodiment of a team that knows their roles: Tony Parker and Tim Duncan handle the midrange, Danny Green bombs away from the wings and the corners, and Thiago Splitter cleans up and shoots almost exclusively near the hoop. There are slight exceptions likely born of player tendencies within the group, but this is a squad with pieces who know exactly what their roles are and remain constantly within them.
Contrast that with the Jazz and the picture is pretty ugly, especially as far as future pieces within last year’s starting lineup go. Favors, as I mentioned above, was valiant in attempting to extend his range, but he needs to take those little clumps and parlay them into larger ones in more concentrated areas. And though my readers must think I hate them after these last few weeks (I don’t), Burke and Hayward were very disappointing here, the former in particular. Trey’s clusters are littered all over the court, clumping heavily only near the basket. Hayward was a tad more selective, but still has splotches in several places and not even a sniff of one or two “go-to” areas.
As a team, these results are especially worrisome within a starting lineup designed by then-coach Ty Corbin specifically to improve Utah’s spacing after a dreadful shooting start[ref]A goal he seemingly accomplished, as evidenced by Williams’ and Jefferson’s charts.[/ref]. Some of it is certainly age, role and lack of experience – Williams and Jefferson are both veterans, and thus have had more time in the league to define their offensive games. But these factors aside, these are still professional basketball players, and both Burke and Hayward count basketball IQ as a positive element in their respective scouting reports. That neither was able to find any semblance of a niche for themselves even among improved spacing can’t be entirely overlooked, and should be one of new head coach Quin Snyder’s first priorities when instilling his system.
Like many elements of a disappointing year in 2013-14, there’s tons of context at play here. This is only a single lineup, albeit far and away Utah’s most-used, and they weren’t exactly playing within a solid offensive system. Also, while the young guards certainly need improvement, a refining of their offensive games is far from out of the question; Burke still has at least a two-year cushion before such a chart would incite truly ominous alarm bells, and the fact that “Hayward was outside his optimal role!” has been repeated ad nauseam doesn’t make it any less true. The NBA is fun in part because ugly situations can turn around so quickly[ref]See “Cavaliers, Cleveland.”[/ref], and despite several decidedly negative elements and their lingering stench, the Jazz have positioned themselves well to undergo just such a reversal should a couple chips fall their way.
Stay tuned later this week for Part 2 of my examination of last year’s Jazz using Matt D’Anna’s TeamSPACE visualization, when I’ll look at Utah’s “Core Five” lineup and forecast their prospects for the upcoming season both as individuals and as a unit.