The best thing about the NBA’s summer league in Las Vegas is that everyone is there — players, scouts, executives, agents, writers and hangers-on. The basketball itself often gets lost amid endless opportunities to chat with team employees and others sitting in the reserved sections of the stands at Cox Pavilion and sitting in random places all over the bigger Thomas & Mack Center.
There were also various consultants hanging around, including a group from STATS LLC, the guys who build and sell the fancy tracking camera systems installed now in 10 NBA arenas — with more teams likely to buy in at a low six-figure price before next season. I’ve written about the data the systems provide several times. The cameras track every movement of player, ball and referees, giving teams that subscribe access to a trove of data — player running speeds, how often particular players hit their highest running speeds (Kevin Love blows everyone away by this measure), how often players dribble, shooting accuracy based on how many times a particular player dribbles the ball before firing away, and a million other things. Teams also hope to learn a bit more about fatigue and health via a news STATS partnership, and one unnamed team already found last season that an early injury to a key starter resulted in the other four starters suddenly exerting a lot more energy — sprinting more, and reaching peak cutting speed much more often.
The challenge for teams is going to be sorting all the data, finding the useful bits and then settling on ways to actually make use of it. Some teams are already ahead of others in that regard.
Over the last year, STATS has been kind enough to give SI.com previews of new data as they get it (and as teams, always so secretive, allow STATS to discuss it publicly), and Brian Kopp, a STATS vice-president, visited courtside with me in Las Vegas for a bit during summer league. He showed me a preview of this must-read charting work by Kirk Goldsberry, currently teaching at Harvard, who used the STATS data to chart whether the location of a shot attempt influences the location of the eventual rebound. (Hint: It does, and more than some prior studies suggested. Go read it).
One fun tidbit Kopp shared: Information on the trajectory of three-point shot attempts for some of the league’s high-volume long-range shooters. This strikes me as the kind of data teams would find relevant, depending on what it shows. Is a guy’s trajectory too high? Too low? To line drive-ish? Does it vary more dramatically than average between makes and misses? This is stuff coaches can use to teach and change behavior; coaches have always worked to teach proper mechanics and shot trajectory, often using sophisticated machines that measure shot arc, but the STATS data provides harder numbers — and more of them from in-game situations.
Findings for every shooter who attempted at least 40 recorded threes show that NBA gunners are just as precise as you’d imagine — with some exceptions. Just about half of all recorded threes — nearly 11,000 shots in all — reached a maximum height somewhere between 15 and 16 feet above the ground. Recorded shooters hit 37.4 percent of attempts that peaked in that range — a couple of ticks above the league’s overall average of 34.9 percent last season. Shooters actually shot just a hair better, 37.9 percent, on shots that peaked between 14 and 15 feet above the floor, but the percentages declined sharply once you got out of that 14-16 foot range. Guys hit only 31.2 percent of threes that peaked between 16 and 17 feet, and percentages got much worse as heights reached extremes on either end.
This isn’t surprising. Line drive shot attempts generally just don’t work, and guys typically add a bit of extra height when pressured — meaning the shot is already relatively low-percentage.
But there are exceptions — proof that guys can scrape the edges of this range and succeed if a certain technique works for them. Ray Allen’s attempts, for instance, reached an average peak height of just 14.69 feet, and he’s probably the greatest three-point shooter in league history. Of all shooters who attempted at least 40 recorded threes, only Kawhi Leonard’s shots had a lower average maximum height. Allen is an outlier, but the trajectory data also provide a clue to his success: Both his makes and misses reached an average peak height of exactly 14.69 feet, a nice bit of consistency and proof that Allen’s borderline obsessive-compulsive work routines (and “borderline” is generous) work well.
Allen’s perfect make/miss alignment is unusual, but not unbelievably so. The average shot height difference in max height between makes and misses was about half an inch, with misses trending higher in the air than makes. Kevin Martin and Chase Budinger also recorded perfect make/miss matches last season, and Danny Green, Linas Kleiza, Kyle Lowry and Kevin Durant both had make/miss gaps of about one-tenth of an inch.
One tidbit about Durant: His attempts reached an average max height of 15.64 inches, one of the half-dozen highest figures in the database. The reason is pretty obvious: He’s tall, with a pretty soft shot. Ersan Ilyaosova had the highest recorded average max height (15.8 inches), a finding that jibes with the eye test; that guy lobs glorious moonballs. And in proving that dizzying heights aren’t just for tall guys, O.J. Mayo (15.7 feet), Brandon Jennings (15.6), Russell Westbrook (15.59) and little Nate Robinson (15.53) rank among the “top ten” in this regard.
Want data on a really wacky three-point shooter? Dive into Jamal Crawford, who had the biggest gap in max height — nearly a full foot — between makes and misses, the largest such gap in the STATS sample size. But in a bit of an unusual twist, Crawford’s makes reached much higher peaks — an Ilyasovian 15.8 feet — than his misses, which topped out at just 15.0 feet on average. Michael Beasley, not exactly a model of consistency in any way, had the second-largest make/miss gap — above five inches.
Is that inconsistency a problem? Maybe. It’s something I’d want to know if I were a coach, and especially the coach of a younger player like Beasley. But there are so many unanswered questions here: Does Beasley’s height data jump around like this among his made shots and not just between his makes and misses? Does Beasley cut his trajectory short when there’s a defender within two feet of him? What about when there’s a defender right in his face? What happens on transition pull-ups, or when Beasley shoots off the dribble? Heck, maybe something funky even happens to a shooter as consistent as Durant when he takes threes off the bounce in a pick-and-roll. Would that even matter?
These are deeper data dives, and the kinds of questions that for now are behind the paid wall — the domain of subscribing teams. But as this sort of data gradually makes its way through all 30 teams, more of it will eventually become public. We will know so much more in 10 years than we do now, and that will create a whole new challenge for executives, scouts, writers and fans: What information matters?