In light of the continued magnifying glass that everyone is using to look at Steph Curry’s individual defense, it makes sense that one of GSOM’s brightest basketball minds would want to dig deeper into the plays behind the numbers.
So while I (this is Duby Dub Dubs writing) pretty much hate the Buzzfeed articles that are just a summary of a Twitter thread, I’m going to do pretty much exactly that: break down Apricot’s Twitter analysis of Curry’s defense in the Golden State Warriors’ 41-point win against the Houston Rockets.
Eric Apricot is simply to busy to be bothered, but I think some of this information is especially germane heading into tonight’s critical game four - so here we go.
Starting off with the premise that maybe the NBA matchup data isn’t great
Here’s what started Apricot’s basketball head gears turning - some of the initial one-vs-one data in the NBA.com data set just looked off:
The isolation stats have been a clusterf. https://t.co/PXnfvxN28q looked way off, who knows where ESPN got theirs. I more trust humans charting off film, but even then how are they counting dump offs etc.— Eric Apricot (@EricApricot) May 16, 2018
I was gonna make a super cut, but I figured I’d wait until after G2. https://t.co/HQANlEijGl
Obviously, these numbers aren’t hard science. It can be incredibly subjective trying to determine whose man was whose. While this is a known problem, it was this article by Sam Amick that first exposed the discrepancy between the official numbers from NBA, and those you get if you just watch the game and tally up the results of the relevant plays.
So? How does the discerning eye of the great Apricot judge Steph?
As my journalism teacher taught me, bottom line up front — generally, Curry was pretty darn good!
Short summary is HOU attacked Steph on 28 possessions and got 32 points (1.14ppp). That's so-so. However, if you count plays where they brought Steph into the action and passed off, that was another 6 poss for 0 pts (and one pick-2). Together, that's 0.94ppp.— Eric Apricot (@EricApricot) May 22, 2018
As with most statistical analyses, it’s important to couch the data within a broader understanding of the significance of these numbers. In Curry’s case, this means isolating the defensive plays that “mattered.” Now, before you all jump down our throats, yes, all plays matter - but anyone who has even passingly glanced at the Warriors this season knows that our focus comes and goes like a bad toupee on a windy day.
Also notable is that HOU stopped attacking Steph when Nick Young and Jordan Bell were on the floor. I didn't analyze them, but the NBA says Bell gave up 15 pts on 10 poss to the main guys, and Young 23 pts on 22 poss. That jibes with my live impression.— Eric Apricot (@EricApricot) May 22, 2018
A couple of salient points here.
First, when the game was still in doubt, Curry was excellent. His final numbers being “fine” are the result of his play later on in the game dragging his metrics down. Maybe this is cherry-picking, but I prefer to think of it as selective analysis.
Secondly, there are some hilarious unintended consequences of Nick Young being on the floor. One of those is that tidbit about the Rockets not attacking Steph as much. In exposing a weaker defender, the Warriors may be inadvertently protecting Curry. Sun Tzu would be proud.
Generally, it looks like the Apricot defensive tally is fairly close to the official NBA numbers - both have Curry as an above-average defender in this series.
Click here for Apricot’s full notes on Curry’s defense in Game 3.