FanPost

Is Sharing Caring Pt. II: The Controlled Edition

Much has been made about the ball distribution of this Celtics team. This year, they've consistently been in the lower middle of the pack in terms of total passes made per game. Regardless of that, the ball movement is good when they win. It's bad when they lose. If I had some super machine learning algorithm, I would love to see if there was a bigger correlation between positive and negative comments in the threads about gameplay (the Celtics didn't move the ball, played bad defense, etc.) and point differential in the game than whether the Celtics actually played well or not, but that's just a pipe dream, as I don't have that kind of data mining skill. I'll leave it to the guys down at Second Spectrum to make this a reality.

Anyway, on to the actual topic, I made this because I was getting reviews from the doubters that correlating passes and offensive rating was never going to work because there are a lot of other factors involved. I did point out that if passes were a significant factor, then it would have showed up somewhat, but the very least, we can say that making passes is significantly less important than actual shot-making or making meaningful passes, like assists. Meaningful passes is the key. Just having volume doesn't make the passes better. Good passes make good shots. More passes just make different ones. So while everyone's been quantifying the volume of passes, I have decided to look at what the passes have actually created.

The easiest and most ultimate form of my analysis would just be using QSQ (quantified shot quality), which is collected by Second Spectrum, but common folk such as myself do not have access to these things. So, I'm going to have to go with the next best thing: defender distance. Second Spectrum and the NBA have been gracious enough to let us see how far away the nearest defender is from every tracked shot. To quantify the value of defender distance, I'll take the averages of the EFG% at each defender distance and subtract the league average EFG% to get a "defender distance bonus" (or DDB) for 0-2, 2-4, 4-6, and 6+ feet between a player and the nearest defender. Then, I will use those to get an aggregate defender distance bonus for a team's shooting profile. Then, from that, I will plot that defender distance bonus against number of passes to see how they contribute to getting better shots.

Long story short: The data shows absolutely nothing.

For the stat geeks, when I plotted the defender distance bonus (which goes from about -0.8 to 0.8) against passes (which ranges from about 250-370), I ended up with a linear fit of 0.0014*X-0.402 with an R^2 of 0.0138 (X being number of passes). For the non stat geeks, that means that the correlation is garbage and even if it wasn't, the relationship is absurdly weak. If a team took 100 more passes (that is, went from being the worst passing team to the best), the trend would expect a difference of 0.14 in DDB, which would move you up maybe 3 or so spots. That means that even if the correlation was better (it's a 0.01 on a 0 to 1 point scale, 1 being perfect correlation), it still wouldn't mean much.

While I'm here, though, with this data, I would like to share some things with the class.

1. The Warriors are absurd. They have a very average shot profile (0.07 defender distance bonus), but they still manage to be one of the best offenses. They simply have better players. Their system doesn't actually generate shots that are that great.

2. Budenholzer is a bit of a genius. The Bucks lead the league in defender distance bonus, even with a superstar who doesn't really shoot well. They get a ton of open looks. Credit the coach. He's running a fantastic offense.

3. Brad Stevens is also pretty good in his own right. The Celtics are 5th in the league in defender distance bonus. We get good looks, but the stats show that we have been awful on our open looks (defender 4-6 feet away) even though we have been stellar with wide open looks (6+ feet). That's going to change.

4. The Pelicans are playing with house money. Far and away, the New Orleans Pelicans, yes, the best offense in the league, has the worst defender distance bonus. Based on their team, I don't expect them to keep this forever. I'm calling for them to have a serious downturn sooner rather than later.

I hoped you enjoyed the article and I'd love to hear what you think!

Credit to Second Spectrum via NBA.com for the passing and shooting data.

FanPosts are fan-created content and do not necessarily reflect the opinions of CelticsBlog. Be respectful and keep it clean. Thanks.