Nylon Calculus: Going to great lengths — height and defensive matchups

OAKLAND, CA - FEBRUARY 06: Raymond Felton
OAKLAND, CA - FEBRUARY 06: Raymond Felton /
facebooktwitterreddit

As with most buzzwords, the term “small ball” oversimplifies a concept that’s rich in nuance. Perhaps the most incisive observation about its “deceiving” quality comes from Steve Kerr.

In discussing a 2015 NBA Finals lineup that featured Stephen Curry, Shaun Livingston, Andre Iguodala, Harrison Barnes and Draymond Green — “all long, active players” — the Golden State Warriors head coach asserted: “That’s not really small.” Apart from Green, who is a nontraditional center, “that’s pretty big at four positions.” The “small ball” label may understandably highlight the most glaring innovation, but it fails to capture all of the other strategic advantages. This shortcoming is among the many reasons why Krishna Narsu prefers to call it “skill ball” instead.

Indeed, for all the talk about undersized yet versatile big men, it’s worth shining a brighter spotlight on rangy yet agile defenders who can use their length to contain shorter wing players. Consider the likes of Andre Roberson, Robert Covington and Luc Mbah a Moute. They hound opposing facilitators and shooters around the perimeter, navigate screens in ways that keep the defense organized, clog up passing lanes, switch seamlessly onto other players and even hold their own in the post. There are also unsung representatives of this archetype who bring similar advantages on the defensive end. For example, Joe Ingles can match up with ball handlers, including Chris Paul, who served as one of his key assignments in last year’s first-round playoff series between the Utah Jazz and the Los Angeles Clippers.

Conventionally, we can examine these defensive dynamics through a combination of box-score, play-by-play and lineup data. For every possession, we know who’s on the court, so we can determine which team has the overall physical edge and even infer whether specific players have a height advantage. Todd Whitehead had a great season preview that leveraged these types of numbers. In The Athletic, he found that defensive units with players of similar size tended to perform well, and the Warriors in particular benefited from having “a glut of interchangeable tall wings” who can switch everything and thwart their offensive counterparts.

We can build upon such analyses with new data from NBA.com. Provided by Second Spectrum, these statistics track individual matchups, allowing us to see the total number of possessions in which a specific defender was guarding a specific offensive player, as well as the resulting box-score figures like points, turnovers, etc.

The data are limited in certain respects. For instance, they don’t indicate which teammates were on the floor during particular matchups, and we’re left to wonder how the matchups came about in the first place (that is, whether the matchups were intended or switched; we do know that, if a defender guards multiple players on a given possession, his “official” matchup is the one in which he spends the most time). Nonetheless, the data provide more granular details than ever before, enabling us to estimate the frequency with which teams enjoy certain advantages.

Let’s start with the scenario that Kerr described above. The following chart presents each team’s total defensive possessions against perimeter players and identifies how frequently its defenders were taller, shorter or the same height as their matchups. Unsurprisingly, the Warriors rank atop the league in height advantage:

Against perimeter players, Golden State’s defenders have a height advantage over 70 percent of the time, with the Los Angeles Lakers serving as the only other team above the two-thirds mark. The Warriors are at a disadvantage only 20 percent of the time. So, for every individual matchup in which their defender is shorter than his counterpart, there are 3.5 possessions in which their defender is taller.

The Dallas Mavericks are on the opposite end of the spectrum. They have a height advantage only about a third of the time, and they’re at a disadvantage more than half of the time. Their ratio is flipped: for every matchup in which their defender is taller than his opponent, there are 1.6 possessions in which their defender is shorter.

We can look at team rosters to identify where the actual height advantages and disadvantages lie. In the following heatmaps, defenders are listed in descending order from tallest to shortest, with their individual matchups broken down by opponent height:

The Warriors’ interchangeability and versatility immediately stand out. They have five players at 6-foot-7, and four of them match up with shorter opponents on a substantial share of their defensive possessions. Nick Young has a height advantage roughly 64 percent of the time, while Klay Thompson, Patrick McCaw and Shaun Livingston are each over 70 percent. Green is the lone player who spends the vast majority of his time on big men.

Similar observations can be made at 6-foot-9. The Warriors have five players at this height, with Kevin Durant (59 percent) and Omri Casspi (66 percent) dedicating a significant number of defensive possessions to shorter opponents. Meanwhile, David West, Kevon Looney and Jordan Bell devote most of their time to taller opponents. The bottom line is that, as Kerr might point out, few perimeter players would use the term “small ball” to describe their personal experiences with Golden State.

On the flip side, the Mavs are often at a disadvantage. J.J. Barea and Yogi Ferrell are 6-foot-0 guards who almost always concede a few inches to their opponents. Dennis Smith Jr., Devin Harris, Gian Clavell and Wesley Matthews are only slightly better, as they each defend shorter players less than a quarter of the time. Overall, for these six players, the weighted average height of their opponents is somewhere around 6-foot-5 to 6-foot-6 — comparable to the numbers for Golden State’s Young, Thompson, McCaw and Livingston despite being shorter than them.

Note that this analysis should be viewed as a preliminary step to what might become a larger body of research on player matchups for the basketball community. There are still many questions left to be answered.

For one thing, I crudely define “perimeter players” here as being 6-foot-8 or shorter, but this cutoff is arbitrary. Better methodologies can be used to arrive at a definition, and broadly speaking, I hesitate to turn continuous variables into categorical ones. Fortunately, in this case, the general order of teams remains the same no matter where the line is drawn.

In addition, there might be better ways to frame the analysis. Player height tends to suffice, but perhaps something like functional height, which Whitehead used in The Athletic to account for standing reach, is more applicable. Position (or some way to define a player’s role on the court) might also lead to improvements, since defensive assignments are influenced by opponent skills, not just by physical attributes.

Next: Nylon Calculus Week 19 in Review -- Sloan Conference!

Ultimately, matchups and results must be linked. NBA.com already provides basic statistics to help measure performance, but further studies are required to separate the signal from the noise. While frequencies are adequate for now, efficiencies have to enter the conversation at some point.