NBA Week 3 in Review: Chasing Trends

With the Cleveland Cavaliers finding their groove and some crazy box scores over the past week, there’s a lot to discuss, but let’s delve into a few other topics and leave the talk of things like Kobe Bryant to the rest of the media.

Doubting Thomas (Again)

Even the smartest people following the league are being duped into a simple correlation/causation error with Isaiah Thomas leaving Sacramento. As I’ve stated repeatedly, Thomas was not the problem with the team last season — they played their best when he was on the court and actually outscored opponents. This is not an aberration either, because we’re seeing that again with Phoenix this season: Thomas has the second highest on/off team numbers behind Markieff Morris. Their offense is exquisite when he plays and he has amazing chemistry with Gerald Green. Phoenix isn’t as strong as they’d like to be because Dragic is struggling and they miss Frye, but that has nothing to do with Thomas and the minutes where they’ve struggled have mostly come with him on the bench, just like last season in Sacramento. (Dragic is playing better without Thomas, but we all knew the dual point guard alignment could lead to problems and it has nothing to do with Sacramento.) So unless you think Thomas has a hex on the other starters when he’s not playing, sabotaging their numbers from the bench, any assessment as Thomas as a player killer who makes everyone worse is off the mark.

What’s happening in Sacramento should be obvious. It’s the same with Davis in New Orleans: a big guy with a lot of talent, coming off a nice experience winning a gold medal in the summer, has a breakthrough season. This was going to happen whether or not they had Darren Collison. The list of modern centers who were as active on offense as Cousins was is an extremely short one and such a high usage rate at a young age is an encouraging indication of star-potential

If Isaiah Thomas was indeed holding Cousins back, we should have seen it in the numbers last season. When Thomas was on the court, via NBAWOWY, Cousins had a 56.1 TS% on a super high 34.2 usage rate with 13.6 turnovers per 100 player possessions. And when Thomas was off the court? The usage was exactly the same with a slightly lower TS%, 55.0, and a slightly lower turnover rate, 12.2 — he was virtually the same. Yet the team had an above average offensive rating when Thomas played at 108.1 and below average at 103.6 when he sat, thanks to a team TS% that took a dive. Digging further into the numbers, and it looks even better for Thomas. Cousins had more shots at the rim with a slightly higher assist rate with Thomas and he generated nearly 50% more free throws.

DeMarcus Cousins, 2014 stats:
With Isaiah Thomas:
Usage% (shot volume):
TS% (shot efficiency):
TOV (turnovers):
Free throws/field goals:
% of shots within 0-3 feet of rim:
% of shots assisted within 0-3 feet of rim:

With Isaiah Thomas:
Usage% (shot volume):
TS% (shot efficiency):
TOV (turnovers):
Free throws/field goals:
% of shots within 0-3 feet of rim:
% of shots assisted within 0-3 feet of rim:

There’s no evidence that Cousins was being “held back” last season and the teams Thomas has been on have only done poorly when he’s sat. Ten games is too early to write off a player and hail another point guard who’s bounced around from team to team because no one’s cared for his results. The Kings are probably playing over their heads, and we can point to a ten game stretch last season where they went 6-4 and beat Miami, Houston, and Portland. Also, the improvement isn’t entirely on offense; half of it is from their defense, which, if it held, would be due to more commitment from Cousins or other players, saying nothing of Thomas.

Roy Hibbert’s Lonely Island

There are few bright spots in Indiana, but Hibbert has valiantly been fighting hard on a ship with no captains. He’s still the master of verticality and probably the best rim protector in the league. Even without Paul George and Stephenson with Hill and West still injured, Indiana has a top ten defense and are defending very well considering all the minutes devoted to guys like Copeland and Scola. He’s a sneaky good pick for the Defensive Player of the Year award too.

But the problem the last two seasons was that his offensive game unraveled and he had problems even finishing at the rim. Thankfully, he’s rediscovered that shooting 67% at the rim, following 2014 and 2013 where he ranked 9th and 8th, respectively in field-goal percentage within five feet of the rim among players with at least 200 attempts. His high jump shot field-goal percentage is probably unsustainable, but he’s getting to the line more than he ever has and, consequently, he’s riding a career high in shooting efficiency while shouldering a proportion of the offense that matches his career average. What matters, however, is that he doesn’t suffer through another Dr. Jekyl and Mr. Hyde season where his first and second half splits are astronomically different. Though so far, his play has been commendable, seen in the clips below.

In the first video, Hibbert pushes the defensive stalwart Gortat deep in the post and gets him off his feet with a nice little up-and-under for a short hook. His renewed vigor has led to more foul shots, seen in the second video where he gets an And-1 versus fellow rim protection master Larry Sanders. And in the last clip, he posts up former Defensive Player of the Year Marc Gasol and hits a hook shot over him.

Defensive Doppelgangers

Checking the player tracking stats for any surprising leaders this season, I was amused by two players with oddly similar stats. I’ll post the stats below (as of Sunday) and the answers at the bottom of the column:

Player A:
29.6 MPG, 1.6 blocks per game, 7.5 Opp FGA defended near rim, 44.0% allowed near rim, 3.1 fouls.

Player B:
27.9 MPG, 1.6 blocks per game, 6.4 Opp FGA defended near rim, 44.4% allowed near rim, 2.6 fouls.

Chase-down Blocks

There are few events as exciting as two super fast players racing in transition against each other, as you can see in the below clip with Shane Larkin and Teague, and it ends with a chase-down block. However, what exactly is a chase-down block? How is it defined? From what I’ve read, the term was invented by Fred McLeod who noticed Tayshaun Prince’s proclivity to swat shots on fast breaks, like his famous block on Reggie Miller. The term was then popularized when Cleveland boasted about LeBron having 23 chase-down blocks in 2009. It entered the national consciousness.

But we are deep into the information age. Shouldn’t this be available somewhere? As long as we agree upon a clear definition, this shouldn’t be too difficult to tally, as long as you have play-by-play data — set standards for time on the shot clock and blocked shot distance. For instance, in the play-by-play log, Larkin misses the three-foot layup five seconds after the possession change. But chase-down implies the defender was behind or at the side of the shooter. Alas, this stat may be lost to the sands of time, though a tally of a number of blocks on fast breaks is still interesting and with SportVU we could still have the stat for the this season and the last.

The Warriors Moving with Purpose

There’s been a lot made about the Warriors and their strong start and many are attributing it to their new offense, but the stylistic changes are hardly the causes yet, sadly. Klay Thompson has blown up, growing his usage rate from 22.6, a bit above average, to a star-esque 28.1. And what’s mot impressive is that his efficiency has skyrocketed too, a rare combination, partly thanks to a new-found ability to draw free throws. (He may have learned these dark arts from James Harden over the summer.)

The Warriors have emphasized ball and player movement, and while their assists are up so are their turnovers. All their fancy passes and set plays haven’t consistently led to better shots. For an example, in the first video below the Warriors run an odd play with three dribble-handoffs in a row in the same spot on the floor, and all it led to was an Iguodala layup at the rim where Brook Lopez waiting for him. There are two other examples below as well showing how heavy player movement and a huge amount of passes led to mere long, low-percentage two-pointers.

Consistency of the SportVU Rim Protection Stats

With only a year of SportVU statistics available, it’s difficult to note which stats are important and stable. A lot has been made of numbers like catch-and-shoot percentages and rim protection metrics, for instance, but we don’t know how much of this is noise and how much is true skill. To account for this, even though it’s early in the season, I looked at every player with at least five games and six defensive attempts at the rim as well as 40 games last season with six attempts. Bridging their 2014 and 2015 stats, we have some idea of how “repeatable” it is.

There are some observations last season that are holding up again like Roy Hibbert being the king of the stat and Robin Lopez’s surprising impact. However, there’s also Zach Randolph sitting at an outstanding 40% allowed and Miles Plumlee suddenly looking terrible. To quantify this, I calculated a simple correlation coefficient — -1 is an inverse relationship, 0 is none, and 1 is a perfect relationship — for opponent field-goal percentage at the rim when the defender is within five feet and the attempts defended against per 36 minutes. The results?

Rim protection FG%:

Rim protection FGA/36 MP:

Given that the p-values for these associated variables are under 0.0001, there’s some confidence here that there’s significance. These aren’t perfect correlations, but given the circumstances this is encouraging. However, what could anyone make of correlation coefficients in a vacuum like this? What makes 0.6 different than, say, 0.4? For a comparison, I ran the same test with the same set of guys for blocks per minute.

Blocks per 36 MP:

Linking this to a study Neil Paine did on how stats hold up year-to year, blocks are one of the most stable stats we have, while correlations near 0.6 are decent but not great. However, this was after only 10 games and some of the biggest misses are due to someone like Pau showing a massive recovery with a defensive minded team and Howard looking as spry as ever. We can have some faith in these numbers, but there’s not an ideal connection to skill. If you player on a super good defensive team, like Zach Randolph or Pau, you will look a lot more effective because your teammates are forcing worse shots everywhere.

We can try to dig and find more gems with these stats, like Portland did with Kaman who rated extremely well last year, with a little more confidence. For example, who are the two players linked to the stats earlier in the article? Player A is Ibaka, annual Defensive Player of the Year candidate, and Player B is rookie Nerlens Noel on the dilapidated 76ers. In many ways it’s more impressive to put up those good rim protection stats without any help. Ibaka’s blocks are nearly half of what they have been in previous seasons, but Noel is crushing him in steals: 0.2 a game to 1.6.

Let’s just hope for Philadelphia’s sake that holds for the rest of the season. They need it.