Something that has been talked about on NBA nerd Twitter recently is why the opponents of the Boston Celtics have consistently missed open threes at a higher rate than other teams over the past decade. This is a pretty weird phenomenon, as you would assume that opponents would make about the same percentage as open 3's across board, but that just isn't the case. Some of this is obviously due to shooting variability, which is a concept I have already talked a lot about in a previous article. While I talked about how single game shooting variance can have a large impact on one game or even an entire season, I don't know if it has a sustained impact over an entire decade skewed so closely to one particular team.
There have been many explanations as to why this may be. The obvious one is that the Celtics just happen to leave the right guys open. On the surface level, this seems to be a reasonable answer. The only problem is that this is the NBA, and every team leaves the right guys open. No team says, "maybe we should let that Curry guy shoot." Coaches are just too good, and players are too smart for any one team to do this so much better than others. Todd Whitehead is the person on Twitter who dismantled the "leaving the right guys open" argument. Another argument could be that the Boston defensive scheme just allows fewer open looks, meaning maybe there is more variability within the shots taken. First of all, that just isn't true when you look at the numbers, but second of all this trend has gone on through multiple coaching regimes. Now that we've eliminated the easiest explanations for why this is, so now we can get into the nerdier stuff.
One suggestion I have seen is that this is because of the background in T.D. Garden. If their backdrop to shooters is completely different in a way that could throw off shooter’s depth perception, then this could be a reasonable explanation behind the shooting numbers. One thing that hasn't changed throughout the years and is unique to the Celtics is the Garden. The only issue with that theory is that this is what the backdrop for shooters looks like:
Perfectly normal. Exactly the same as the other 29 arenas. So now that the backdrop theory is ruled out, I want to talk about a really nerdy explanation. I've seen people suggest that the difference in opponent 3-point shooting could be because of the angle of the tracking data cameras in T.D. Garden. The way tracking data works is that there are six cameras in every arena that record X, Y coordinates of every player on the court hundreds of times per second. The cameras look at the middle of the torso of the player for their spot on the court, and if this is off because the camera angles are off then it could make a difference. The issue I have with this theory is the question of why the angles would be different and why the NBA wouldn't fix it if they were giving them faulty numbers. I think that this is an interesting theory, but I don't really buy it. Another theory I have heard related to this is that the tracking data is limited to the players torso instead of their hands, so if a long player is contesting a shot from 7 feet then they might have more impact. The problem with this theory is that if I think about the players the Celtics have had over the past 10 years, I don't think they have had an inordinate amount super long players.
Now that I've talked about and eliminated all the theories that are out there, I wanted to dive into something I heard on the Thinking Basketball podcast. Host Ben Taylor gave an interesting note about how sometimes free throw and three-point percentage can be lower against really good players because they force the defense to work more, which then tires them out. After hearing that, I thought that maybe there would be a similar impact when playing against a team that runs more than others on offense. Different teams each have their own offensive scheme, and some have more player movement than others. I wanted to look at if team that have players move faster or longer on offense have an impact on opponent shooting numbers.
I decided I wanted to use the opponent data on NBA.com for wide open 3-point percentage (6+ feet of separation) and free throw percentage. With that data, I would compare it to the average speed and distance from each team while they are on offense. Here are the results:
Yep. The exact opposite of what I thought would be the case is the result of this. That happens sometimes, but it does seem strange that opponent free throw and open 3-point percentage both increase as offensive speed and distance traveled increase. I have a couple of theories for why this may be. First of all, I did not adjust the numbers for who was shooting the open shots or who was shooting the free throws, so that is still a possible factor. My next theory is that maybe when teams run a lot on offense, it can actually make them tired, which can lead to defensive breakdowns (leaving better shooters open) or lazier closeouts. My final theory is that maybe this is all just noise from single season shooting numbers and more data is needed to make a conclusion. Whatever the case, the Boston Celtics Jedi mind trick 3-point defense still remains one of the big mysteries of NBA nerd Twitter.
The code for the project can be found below:
Comments