By Krishna Narsu
The Golden State Warriors have dominated in the playoffs so much that finding any potential weaknesses’ seems to be an exercise in futility. But is their 3-point defense secretly their weakness? Or, at least, a relative weakness, especially compared to previous years? Let’s take a look.
The best way to defend the three is to avoid it altogether (there’s a higher correlation year-to-year or half-year to half-year for 3-point attempt rate allowed versus 3-point percentage allowed). The Warriors have done a very good job of this in the past but during this season, they slipped a bit:
|Warriors 3PA rate allowed|
However, when we look at how they’ve done in the playoffs, it appears that the Warriors have managed to go back to their normal ways of deterring 3-pointers:
|Warriors 3PA rate allowed|
But those numbers may be a bit deceiving. When we break down the Warriors 3-point attempt rate allowed based on each series, we see the Warriors did a significantly better job against the San Antonio Spurs than the Utah Jazz and Portland Trail Blazers:
Before the series against the Spurs, the Warriors actually ranked 11th in 3-point attempt rate allowed. And when we look at the Spurs series, an interesting trend emerged:
Once Kawhi went down, the Warriors did a significantly better job of deterring 3-pointers (the Spurs also missed more). Of course, Kawhi only played 24 minutes in that series and so the numbers with Kawhi on the court are a pretty small sample, but the higher 3-point attempt rates allowed against the Jazz and Blazers suggest that the Warriors 3-point defense might not have actually been better in the playoffs and that they were just fortunate to face a Kawhi-less Spurs offense.
So is this the whole story? Is the Warriors 3-point defense really a weakness? While deterring threes is always the preferred way to measure 3-point defense, there is a small correlation between teams who deter 3-pointers and allow a low 3-point percentage on the threes that they do allow. And the Warriors are one of the teams who’ve shown a consistent trend in allowing a low percentage on 3s:
|3FG%||FG% rank||3FG%||FG% rank|
The Warriors have considerable length, which as I showed here, can have an impact on lowering the opposing teams’ 3-point percentage. Just look at the Warriors players’ wingspan relative to their position average (data from draftexpress.com):
All but one player (Curry) has a wingspan above their position average, and even Curry is basically average for the typical PG. So when factoring in the Warriors considerable length, it’s not surprising that they’re consistently among one of the best teams in opponent 3-point percentage allowed.
So does this mean the Warriors 3-point defense is, in fact, good? Let’s consider one last factor: the types of threes they allow. Using my metric KOBE, which is a measure of shot quality that factors in shot distance, defender distance, catch and shoot versus off the dribble, shot clock and the height differential between the shooter and defender, we can see the quality of 3-point looks the Warriors are giving up.
|Warriors Defensive KOBE on 3s||Playoffs|
In general, the Warriors are typically forcing difficult threes. However, this postseason has been different. Their defensive KOBE on threes through the first two rounds was 11th out of the 16 teams in the playoffs. We can see a series breakdown as well:
|KOBE on 3s|
As we saw earlier with regards to their ability to deter threes against the Jazz and Blazers, they did not do a good job in either of those series. Unfortunately, I don’t have KOBE data for the series against the Spurs but we can look at how they’ve done in the entire playoffs at limiting open threes as well as catch and shoot threes.
|Off Dribble 3FG%||24.78%||3|
As we see here, the results are a bit of a mixed bag. They’ve allowed the highest percentage of open 3-pointers in the playoffs, which, combined with their average ability to deter threes, is not a good sign for them going forward against the Cavaliers, who will launch a ton and mostly make them. However, they are above average in forcing teams to shoot more of the off the dribble variety of threes. This isn’t too surprising when you consider their length and ability to close out.
And as we saw before, the Warriors are great in all forms of 3FG% allowed. Now, is some of that luck, specifically with regards to their Open 3FG% allowed? Or perhaps they’re leaving the right shooters open? While the evidence in the regular season suggests that the elite defenses don’t choose which shooters to leave open any more than any other defense, the playoffs are a different animal and we generally see more strategy employed in a playoff series.
So how can we determine this? We can look at the weighted average percentage of all of the players who’ve taken a 3-point shot against the Warriors in the playoffs. And because 3-point percentage is so highly variable from one season to another, we’ll use career numbers as well as career FT% (a player with a higher career FT% should theoretically be a better shooter).
|Career 3% and FT% of all players who took a 3-point shot vs. Warriors Defense|
|Career 3%||Career FT%|
As we can see here, the Warriors have actually allowed better 3-point shooters to take shots against their defense. So there doesn’t seem to be any evidence that they’re leaving the right guys open. In particular, Hayward actually had a higher 3-point attempt rate against the Warriors than he did in the regular season (40 percent vs. Warriors, compared to 32.4 percent in the regular season).
So while the Warriors length can at least explain some of their low 3FG% allowed, it does appear like they are getting a bit lucky as well.
As a final examination of the Warriors 3-point defense, let’s turn to the tape. I watched every 3-point shot they allowed vs the Jazz and the first game against the Spurs (after that, the Spurs were clearly overmatched and, as we saw earlier, their 3-point attempt rate went down a ton after Kawhi got injured. As it is, they don’t normally shoot a ton of threes.) Of the 154 3-pointers they allowed in five games, I counted approximately 28 where the Warriors played bad defense on a good shooter (if the Warriors allowed a bad 3-point shooter to take a 3-pointer, I ignored it). Let’s look at a few examples:
In this example, we see the Warriors fail to pick up a trailing Hayward in transition:
Here’s another example where the Warriors give Hayward a good look in semi-transition. Durant, for some reason, decides to defend Mack instead of sticking with Hayward. And while he might’ve been expecting someone to pick up Hayward, why leave the Jazz best player?
Here, we see Draymond Green over-helping. He goes for a steal here instead of making sure to stick with Hayward:
Here we see Green over-help again, which was something he did a lot of in Game 1 against the Jazz. This wasn’t really an issue after that though.
Here we see Curry, who is in position, not even bother to contest:
Here we see Klay Thompson over-help, though with the threat of Gobert, you can at least understand. However, he never even bothers to try closing out:
Here we see Zaza Pachulia over-help for way too long, which leads to an open Joe Johnson 3-pointers.
For the most part, the Warriors were pretty solid with their 3-point defense in Game 1 against the Spurs. However, here’s an instance where Durant takes a bad angle which leads to a good look for Kawhi:
Of course, it wouldn’t be nice to focus solely on the negative. The Warriors’ length, specifically Durant here, allows them to make insane plays such as this:
Watch a Warriors game and you can see Durant’s insane closing ability; and with above average length everywhere, the Warriors in general are very good at closing out on shooters.
However, as with all defenses, there are mistakes. For Cleveland to have any chance in this series, they’ll have to take advantage of every Warriors’ mistake. Given the Cavaliers’ fantastic shooting, the Warriors will need to cut down on the number of open 3-pointers they’re allowing, even if their length does make it difficult.
Wingspan data from draftexpress.com. KOBE data from SportVU data. All other data from basketball-reference.com or nba.com