Running Some Numbers At The Break
By Seth Partnow
Feb 12, 2015; Chicago, IL, USA; Cleveland Cavaliers guard Kyrie Irving (2) dribbles the ball against Chicago Bulls guard Derrick Rose (1) during the second half at the United Center. The Chicago Bulls defeat the Cleveland Cavaliers 113-98. Mandatory Credit: Mike DiNovo-USA TODAY Sports
Over the 1230 games that make up an NBA regular season, a veritable avalanche of numbers and statistics are generated. While we at Nylon Calculus obviously think this is a good thing, it can be a bit much to remember them all. It’s all too easy to “fire and forget” a piece of analysis and never revisit the findings to see if an observation still holds or there is something to be learned between changes in the results over time. In that spirit of retrospection, some quick updates on some of the trends and topics from the first half of the 2014-15 season.
The Importance of Effective Ball Movement
A major reason why many expected an improvement in Golden State from last year was the likelihood they would better take advantage of their teamwide playmaking skills. Last season’s team featured relatively little ball-movement, perhaps overly reliant on former Coach Mark Jackson’s desire to attack perceived mismatches. Somewhat as a result, the Warriors finished last season as the 12th most efficient offense as measured by TNC’s true possession based calculations. This year, they are up to third. It’s simplistic to ascribe all of this uptick to the change in coaching philosophy from Jackson to Steve Kerr. Certainly Klay Thompson’s improvement from complimentary piece to true Three-and-D All-Star as well as Marreese Speights season-long imitation of a flame-thrower have helped. But it is inescapable the Dubs are moving the ball more[1. 313.6 this season, 11th in the NBA, up from a distant last place 245.8 last year.] and far more effectively — the Warriors lead the NBA is assist chance per game at 50.5, per SportVU.
Unsurprisingly, marrying this improved ball-movement with the already-present offensive talent has seen an explosion in productivity, as Golden State lead’s the NBA in shooting at the break, with the similarly pass-happy Hawks and Clippers rounding out the top three shooting teams:
As was the case last season, every team in the league is shooting much better on assisted shots than on unassisted, though there is fairly wide variation among teams:
There’s a lot to tease out there: Philly has no offensive talent. Oklahoma City’s relatively low differential might mean something about their offensive structure, or could just represent having played half the season with no Kevin Durant. Teams which offensive rebound well have narrower splits than those who don’t.[3. Not shocking, considering how efficient offensive rebound putbacks are relative to “designed” half-court offense.]
Unsustainable Starts
In late November, I looked at the starts to the season experienced by Toronto offensively and Houston defensively and concluded that based on the shots they were taking and giving up respectively, those starts were at least mildly illusory. In particular, Houston’s “Jedi mind trick” three-point defense seemed ripe for a correction. And to some extent, that’s happened. The Rockets’ eFG allowed has gone from third in the league to 11th as their 3FG% allowed has jumped from 26.8% to 31.6% on the season. While still very good, the 33.1% Houston has allowed since that original post is no longer exceptional. This change is opponents’ shooting has by-and-large been caused by opponents simply making more shots. Though giving up about one extra wide open three every other game from the start of the year, the Rockets are seeing more shots go in against them with defenders close (under 4 feet)[4. 26.5% accuracy on the season as compared with 19.3% early in the year.], moderate distances (4-6 feet)[5. 31.6% on the year vs. 25.6% to start.] and not close at all (over 6 feet)[6. 33.8% on the season vs. 31.9% early.] Houston is still playing good defense, especially with Dwight Howard in and out of the lineup[7. And not really up to his usual standards in terms of rim defense when playing.], so this drop represents a classic regression to the mean more than anything else – the Rockets were on the good side of defensive fortune for a while, and have been less so since.
Similarly, early in the season Toronto was hitting a ridiculous array of shots, making 50% of their well-guarded three attempts. Considering the league as a whole shoots under 30% on those shots, that was a noteworthy bonus. This shooting has…not held up as Toronto is down to 33.5% on those same shots, meaning they’ve hit a much-closer-to-the-norm 28.4% of these tightly defenderd shots.
Point Guard Personalities
This being the second season with data available, looking to see if and how the Point Guard Personality Test holds up as players change teams will continue to be of interest at least to me. For the most part, players in big minute roles appear fairly stable season to season, while backups and individuals suddenly playing in radically different context have seen their numbers change. Kelly Scaletta recently broke down the changes in Kyrie Irving’s game from being the unquestioned[5. By everyone except Dion Waiters.] Alpha in Cleveland to a more complimentary role alongside LeBron James within the framework of the PGPT. That analysis is worth a read in it’s own right, and caused me to produce this quick comparison:
Basically, Irving has the ball a fair amount less, but is looking to drive to the basket a lot more often this season.
Another quick comparison is of two players with similarly scouting reports coming in. Orlando’s Elfrid Payton and Boston’s Marcus Smart where both advertised as great defenders with the ability to get into the lane and questionable jump shots. Their defensive reps have proven largely justified, but strangely their offensive games have been massively divergent. Payton can’t shoot a lick, but Smart has turned into one of the most extreme spot up PGs in the league in his first season:
In fact among the 80-some players who have gotten significant minutes at the point in the NBA this season, Smart has gotten to the basket a the third lowest percentage of his chances[8. Drive chances are front court perimeter touches where the player doesn’t catch & shoot.]:
The average for all PGs is around 10%. The lack of driving isn’t necessarily a good or bad thing – Jeff Teague is among the league leaders at 16.7% while Chris Paul just missed the above list at 6.8%, and both were worthy All-Stars. But it is mildly surprising to see Smart’s (and Dante Exum’s to a degree) game translate to the NBA this way.
I’ll have a longer look next week at the developments and changes in the PGPT scores of players around the league, but that’s a quick look for now.
Updated Metrics
In addition to the items above, I’ve updated many of our season-long metrics, including Rim Protection, True Usage, Expected Field Goal Percentage (both team and individual) as well as Kevin Ferrigan’s “Daily RAPM Estimate” or DRE[10. Don’t forget about it.], check em out.