This past weekend, I attended the annual MIT Sloan Sports Analytics Conference (SSAC) in Boston, Massachusetts. This was my third time going to the SSAC, and it is always a great experience. The SSAC, for that weekend, was a mecca of some of the best analytics minds in sports ... but it was also a gathering of sports power brokers, including team owners, general managers, coaches, former players and analysts.
Put all of that expertise together, and you're bound to run into some really cool conversations. Today, we'll look deeper at a conversation that took place in a panel called "Take That for Data" that featured Houston Rockets GM Daryl Morey, ESPN's Jalen Rose, ESPN writer Zach Lowe and Boston Celtics front office honcho Mike Zarren:
Next up: basketball #Analytics panel #NBA with @dmorey @JalenRose @ZachLowe_NBA and @mikezarren moderated by @getnickwright #SSAC18 pic.twitter.com/BalJsFFu33
- Andre Snellings (@ProfessorDrz) February 24, 2018
This was a cool panel, because there was a nice mix between two front-office types who both are strong at analytics, an analyst in Lowe who also is excellent with the numbers, and a former player in Rose, who has some facility with the numbers but more valuable experience as a former player.
Among the many topics they discussed, one thing that they hit upon was that the NBA currently features two of the most efficient offenses in NBA history in the Rockets (116.0 points scored per 100 possessions) and the Golden State Warriors (115.4 points scored per 100 possessions). According to the panel, the Rockets are currently the most efficient offense in NBA history and the Warriors are in the top 10.
The stereotype for this era is that offenses are about only 3-pointers these days, with the basketball trey compared to the baseball home run as one of the more exciting plays in sports. Stephen Curry, more than any other player, is responsible for bringing that joy of excellent long-distance shooting to the masses and making it a fan favorite.
However, it is Morey, more than any other general manager, who is credited with digging into the analytics and deciding that spamming 3-pointers is a more efficient team offensive approach than building an offense around mid-range jumpers. And, to take it further, it's Rockets coach Mike D'Antoni who helped engineer the "seven seconds or less" offense last decade that first made the team 3-point barrage into an unapologetically vicious weapon.
Thus, it makes since that the Rockets and Warriors would be poster children for this new wave of hyper-efficient offenses, built around the strengths of the 3-point shot. Clearly, the two teams have figured out how to get the most bang out of treys better than anyone else, and that's why their offenses are so similarly and historically deadly. Right?
Not so fast.
Because, as the panel pointed out, the Rockets' and the Warriors' offenses are actually nothing alike -- as in not at all. They may end up with a similar destination, as far as effectiveness goes, but their pathways to get there are entirely different.
Let's examine:
Team 3-point percentage:
Warriors: 39.6 3FG% (first place in NBA)
Rockets: 36.7 3FG% (10th)
Team 3-pointers attempted:
Rockets: 42.6 3-point attempts per game (way first place in NBA, second place is 34.9 per game)
Warriors: 29.9 attempts/game (11th)
These stats perfectly reflect team personalities. The Warriors don't shoot that many 3-pointers, but they make them at the best clip in the NBA. The Rockets, on the other hand, just shoot way more treys than anyone else. They do so at a reasonable clip, but they are more about volume than accuracy in taking advantage from downtown.
Let's take a look at another couple of stats that illustrate further areas of offensive difference:
Team assists
Warriors: 30.6 assists/game (way first place in NBA, second place is 26.3 assists/game)
Rockets: 21.8 assists/game (23rd)
Team free throws attempted:
Rockets: 26.5 FTA/Game (second)
Warriors: 21.1 FTA/Game (17th)
This dichotomy is a due to the type of offenses that both teams run, which cater to the strengths of their best players. The Warriors run a free-flowing, motion offense where the ball is zipped around on the pass and ends up in the hands of their best shooters. The offense is set up to lead to open treys, so there aren't a lot of shots at the rim that would lead to free throws.
The Rockets, on the other hand, put the ball in the hands of one central ball-handler, who essentially plays iso or pick-and-roll on every play, creating a lot of shots at the rim and tons of free throws. If there's no opportunity to finish at the rim, then the defense has typically been deformed by the penetration enough to lead to wide open kick-out 3-pointers from their role players.
There-in lies another difference to consider: the distribution of 3-point attempts. Of the Warriors' 29.9 3-point attempts per game, 26.8 of them come from their big four of Curry (10.1 3FGA), Klay Thompson (7.1 3FGA), Kevin Durant (6.0 3FGA) and Draymond Green (3.6 3FGA). Meanwhile, though the Rockets do average 10.6 3FGA from James Harden and 6.6 3FGA from Chris Paul, the majority of their 42.6 3-point attempts come from the non-creators in the offense.
This, then, leads to the fantasy outcome of this particular difference in approaches.
For the Dubs, most of the shooting is done among the stars. They've got three of the best shooters in NBA history starting on the perimeter, a functional long-range shooter in a big-man slot, and their offense is built to put those shooters in positions to get easy looks. However, the offense adapts when the shooters are out of the lineup. In other words, the Warriors' offense isn't set up to get the point guard an open trey, it's set up to get Steph Curry an open trey.
If Curry is injured, then backup Shaun Livingston isn't going to come close to (or even attempt to) do the things that Curry does, even at an attenuated rate. Livingston is a role player, a defensive player who can run the offense. So if Curry goes down, the 3-pointers will funnel to the rest of the "big four." Same if Andre Iguodala replaces Durant, or if Jordan Bell replaces Green. Neither are shooters, and the Dubs will funnel the shots elsewhere. Now, Nick Young will get up his shots if he replaces Thompson, but other than that, there's not much handcuff value for the Warriors' bench players if the starters get hurt.
The Rockets are different. Because most of their treys come from their non-lead guard, in case of injury, there is always more production to be had. So if either Harden or Paul go down, or if any of the other starting non-centers go down, the replacement starters could immediately become fantasy relevant because the replacement is going to be expected to come in and take a bunch of 3-pointers.
We saw that when Gerald Green came in to replace Harden and suddenly was knocking down loads of 3-pointers and scoring 20 points a night. In fact, the Rockets have a whopping 12 players on their roster who average more than five 3-point attempts per 100 possessions. That is almost absurd but guarantees that as long as one of Paul or Harden are healthy and able to run the ship, the Rockets are full of handcuffs who can come in and produce in place of their injured stars.
The Rockets and Warriors are two of the greatest offenses in NBA history, but they do it entirely differently. This made for excellent conversation at the Sloan Sports Analytics Conference -- and it also makes for interesting fantasy outcomes to consider.
