While most fans have now turned their attention to the bright lights of the conference finals, I've spent the last couple of days kicking through the ashes of the opposition from the first two rounds.
While sifting through data trying to dig up something worth studying I decided a nice starting place would be to compare and contrast the first two rounds this year (statistically) to the previous post-season.
Which lead me to create this table:
At the risk of sounding a little condescending, it should be clear to most fans that this Warriors team is a bit of an anomaly. Statistical output aside, there hasn't been many (if any) teams that have managed to accumulate this much talent while preserving the "team first" ethos that allowed them to blossom from the depths of basketball mediocrity and helped turn them into contenders in the first place.
Statistically, however, it's a bit of a mixed bag. We start with the bog standard box score numbers; steals per game, for example, hasn't changed much. The Warriors have been consistently great at forcing turnovers and intercepting passes for a while now and that isn't changing anytime soon. Curiously enough, despite the talking heads on television telling us the Warriors now lacked rim protection after letting go of Andrew Bogut and Festus Ezeli, they're averaging significantly more blocks per game while their defensive field goal percentage and defensive rating have also improved over last season.
A couple of statistics I didn't include in this table are deflections and loose balls recovered. The Warriors want to win; nothing paints a clearer picture of that fact than these two hustle stats. In 2016 during the first two rounds, the Golden State Warriors averaged 13.2 deflections and 6.2 loose balls recovered; this year in the first two rounds they have averaged 18.6 deflections and 12.9 loose balls recovered. This is a result of two things; first of all, they really want to win every single possession, and secondly, they have a lot of "length" on the team. Sure they might have sacrificed traditional size — to an extent — when trading away their "true" big men. But the addition of players like JaVale McGee and Kevin Durant has only added to the never-ending wingspan of this roster (Shaun Livingston, Draymond Green, Klay Thompson etc are all relatively "lengthy" players).
Peaks and troughs
As I continued to explore the gaudy statistics that the Warriors put up during their early playoff run a concerning pattern emerged. I wanted to examine the Warriors performance in the first two rounds on a quarter by quarter basis, and while there are some stunning examples of peak level Warriors doing things only they could, there are also a handful of statistics that left me scrambling for answers.
Above is a graph plotting the Warriors average offensive rating throughout each quarter of the first two rounds.
Clearly, something happens that causes the offense to wilt, going from historically spectacular in the first quarter to simply "okay" in the second quarter.
Fortunately, it's not that much of a mystery; the lineup that plays the most minutes in the second quarter is:
Ian Clark - Klay Thompson - Andre Iguodala - Draymond Green - David West
This lineup isn't bad offensively; it's simply that it isn't really built to trade shots with the opponent. Klay Thompson has been struggling offensively throughout the playoffs while West and Iguodala aren't the offensive players that they once were. Clark is one hell of a spark normally but he's prone to tunnel vision.
Should we be worried about this? My first instinct was not really. I mean even if this lineup isn't amazing offensively they seem to be built more with defense in mind right? There's some good wingspan here and some very intelligent defenders in Iguodala and Green.
I need to pause for a moment and state something that's been on my mind since I started writing about basketball analytics; discovering something you didn't expect and can't fully explain is both fiercely confusing and weirdly exhilarating.
In the chart above we are looking at the Warriors’ average defensive rating throughout each quarter in the first two rounds. To state the obvious because this is defensive rating, the lower the number is the better the defense, meaning that giant spike in the second quarter isn't a good thing, actually; it's a little unnerving.
The Warriors’ average defensive rating in the first quarter is 80.1. This is astounding, greater than great.
In the second quarter that defense inexplicably falls off a cliff. It's the type of defense that should only exist in Ron Adams’ nightmares.
So what's actually going on? Why is there such a significant drop off?
Take that for data!
To begin with, the two most used lineups in the second quarter during these first two rounds were:
Clark - Thompson - Iguodala - Green - West
Curry - Thompson - Iguodala - Durant - Green
The first lineup eats up the vast majority of the minutes. As I said before on paper, this should be a pretty solid defensive lineup; at least three out of five of those players are well above average defensively. In fact, Iguodala and Green are arguably two of the best defenders at their respective positions.
The second lineup every Warriors fan should recognise; it's the infamous "death lineup.” The same lineup that's supposed to obliterate opponents.
Initially, I had no idea how to explain this. So I dove deeper into the numbers and took a look at each individual and how they were performing defensively. Within five feet of the rim, Draymond Green (a top candidate for defensive player of the year) allowed opponents to shoot 67.3% while David West allowed 75% and Kevin Durant allowed 73%.
The defensive rating of Green is an abysmal 126, and for Thompson and Iguodala, it's 127 and 132, respectfully.
Meanwhile, Zaza Pachulia has a defensive rating in second quarters of 85.4 and only allows opponents to shoot 25% from within five feet of the rim. Could it be that Pachulia is the fulcrum of the Warriors defense? Well... probably not. It's possible they're feeling the effects of a lack of "size" but given the amount of switching and team-oriented style of defense I find this hard to believe.
After talking it through with Eric Apricot (the superb mind behind GSoM’s "Explain One Play" articles) we arrived at a possible explanation.
It's feasible that Thompson’s shooting woes, combined with the streaky offense of that specific unit, is having a negative effect on the defense. As Apricot put it himself, “If the offense is bad you get counterattacks, early offense and sometimes fast breaks.”
The numbers back this theory up. The Warriors allowed an average 2.2 fast break points through the other three-quarters and an average 2.6 in the second quarter. This jump might seem negligible to some but the other statistics are even more telling. The average points off turnovers the Warriors allowed in the other three-quarters is 2.4 per game while in the second quarter that number shoots up to 5.5. In other words, on average, the Warriors allowed more than double the amount of points off turnovers in the second quarter as they did in the other three combined.
Furthermore, the points allowed in the paint increased from an average of 8.6 in the other three-quarters to 14 in the second quarter.
As Apricot said himself, the struggles on offense only amplify the carelessness on defense during this quarter.
Regardless, I do feel there are a couple of things that I should mention here. To begin with, general fatigue could be playing into this to an extent but more importantly, this is a very small sample size. In reality, it's probably not something fans should be concerned about, at least, not yet.
Overall the Warriors have had a fantastic post season. The numbers through the first two rounds back this up for the most part and if you've been watching the games, their level of play has clearly been head and shoulders above the rest.
The first two rounds can be summarized quite easily; opponents do what they can while the Warriors do what they want.
[Editor’s Note: All stats taken from stats.nba.com]