At this point in the season if you haven’t picked up on my new found fascination with shot location data then I should welcome you to the blog. Welcome! But as different shots hold different values, and different players different skill sets, I wanted to learn how and where different players are affecting a basketball game. This is the story of how I got to asking the question.
The transit between my first two BART stops on the way home from work gets no cellular reception. If I’m able to get a Twitter refresh before frantically boarding, it means I scour over the last few hours of tweets. That night I came across this:
In a further panic than the aforementioned refresh, I managed to email myself that tweet as it had piqued my interest. I then went home, ordered $45 of Indian food and watched Elysium with my brother. He passed out while I grabbed my laptop. I was setting out to answer Nieves’ question. Where does a specific player (Kyle Anderson in this case) most effectively and frequently affect a basketball game?
We first needed to know how often Kyle even had a chance to affect the game. One component of this would be to look at a player’s %poss or usage rate. This tells us how often a player is shooting, passing, turning over. It’s a great number but without context it just shows us where the ballhogs are (for better or worse). KenPom often marries usage with ORtg to see if players are being efficient with the possessions they get. It’s a far better marriage than anything Kardashian but not good enough for Neal’s answer.
I needed to know how much of Kyle’s usage was coming at the rim or otherwise. From the hoop-math, I can tell you Kyle’s FG%, percent of shots and assists at the rim, on 2-point jumpers, and from three. I could have told Neal some of these numbers and perhaps satisfied his question. But I wanted to answer it. I knew we’d have to marry up KenPom and hoop-math.
From Pomeroy we could capture Kyle’s possession data and from hoop-math his location data. Our first calculation was to understand how often he had the ball, uncovering how many possessions Kyle was involved in per game. It’s a complex stat but after consult of greater minds than my own, we agreed that the following would suffice in ball parking Kyle (or any player’s) individual possessions per game:
Individual Possessions/game = (team possessions per game)*(%min)*(%poss)
This is taking into account the total number of possessions a player’s team is getting per game, the percentage of minutes he’s on the floor for those possessions, and the percentage of possessions he’s involved in. With this number we understand approximately how many shots, assists, and turnovers Kyle is a part of. More visually:
Slow-Mo gets the third most possessions per game amongst ‘participants’ in my initial study involving the Pac-12 elite (only Roberto Nelson and Jahii Carson were getting more touches per contest). With the knowledge of how often Anderson was doing something, it was time to discover where he was doing it.
Here is where I have to tell you that our final answer is going to be inexact. That’s ok, right? A study like this is a fun examination into that great middle ground between perception and reality. Without Synergy Sports
I’m not soon going to look at all of Kyle’s possessions to discover how often he is indeed getting into the paint and scoring or dishing. But the perception is that he’s doing it often; it’s why Nieves asked the question and I imagine you’ve noticed it too because I know your team’s struggled to stop it. What we’re figuring out here is approximately
how often Kyle Anderson is helping his team be successful from inside the arc. Kapeesh?
And so how often are those 15-ish possessions resulting in something around the rim?
||% poss resulting in rim score
||% poss resulting in play at rim
||Success % at Rim
So more than 40% of the time Kyle Anderson is involved in a play, it results in something happening at the rim. And on 34.51% of Anderson’s possessions, someone in powder blue (Anderson or otherwise) is scoring at the rim. Stand alone numbers are rarely significant but let me tell you something: that’s significant. I’ve exhausted rim data on the blog but if the average FG% at the rim is 61% then it would seem to behoove your team to shoot there. Kyle Anderson ensure that it happens more than 40% of the time!It was the fourth highest percentage of possessions in the study but the third most scores at the rim per game (5.16) against the fifth most plays there (6.4).
And he’s not just flailing in there, diving recklessly into the paint with no where to go. Looking at the difference between his scores and plays at the rim (those last two numbers from the paragraph above), we find that 80.63% of his rim possessions are resulting in two for the blue. That’s the fourth best percentage amongst the guys I studied. By comparison, Jahii Carson has 40.06% of his possessions at the rim (shot taken or assist made) but only 62.05% of those result in a rim score (more on Carson later this week).
Here is the full table ranked by success at the rim:
||% poss resulting in rim score
||% poss resulting in play at rim
||Success % at Rim
|Brett Comer (2013)
This is a ranking of effectiveness when making plays at rim (third column). The players chosen was essentially arbitrary and ad-hoc based on who I thought was driving and dishing. Email or tweet me if you want me to get your guy. There are infinite depths by which to dive further into this and I intend to do that on a team-by-team basis over the coming weeks. But above is a Pac-12 snap shot.
Oh, and you’ll notice Brett Comer. Brett was Andy Enfield’s Florida Gulf Coast point guard and I figured the leader of Dunk City would be an interesting study. Turns out I was right as he CRUSHES the Pac-12 guys in percentage of plays at the rim (72.47%). Anyhow, more on that later.
Back to your question, Neal. Kyle Anderson is creating a play at the rim 6.4 times per game and turning 5.16 of those into a score or assist. Only Delon Wright creates more scores at the rim and he’s a freak (7.08). But as the question asked about Kyle’s time in the paint, we could also include the two-point data. I was hesitant to do such considering that’s a much bigger and less effective shooting range. Plus, the Wear twins love shooting twos with their foot on the arc and ain’t nobody but evidently Kyle Anderson got time for that.
As it were, amongst those studied, he creates the second highest percentage of scores from 2-point range (3.92/game, 25.5% of his possessions). He’s the fourth most successful at converting these plays to points (54.75%). Once again, Anderson is setting his teammates up to be successful, but like I said, the two-point data doesn’t fully scratch the itch. The rimformation answers Nieves’ trigger question.
But to examine Anderson as a complete game affecting package, we had to see what he did in creating threes, too.
Anderson blows the rest of these players out of the water when it comes to effectively creating three pointers. Of his possessions that result in a trey, 78.95% of them are successful. Next best in conference is Jahii Carson’s 69.93% (which is why I believe he has such a low rim success rate but, again, more on him later). And this isn’t even a diluted stat. Anderson creates the second most three point scores per game (3).
So between the rim and three data, I’m drawn to three conclusions: 1) Kyle has a great ability to draw multiple defenders to himself, 2) he’s very adept at finding the man left or the hole created, and 3) his own shooting. Slow-Mo in an efficient 22-42 from distance this year and so he’s either shooting a good three or passing to one.
To summarize (again): 80% of Kyle Anderson possessions that result in a rim shot or three point shot go in the hoop. Kyle’s helping you help him.
We’ve long known Anderson was a terrific basketball player and we’ve long called him a unique talent. As we said early on, this was a look at the balance between perception and reality. We assumed this about Kyle, we’ve proven this about Kyle. Only he and TJ McConnell rank in the top four of all three locations by way of success percentage and he’s the only one who is 6’9″ within that two-man subset. Unique to be certain.
I’m going to use this information to hopefully learn a little more – and share -about each of your teams. As I mentioned, I found the Dunk City stuff to be fascinating as well as the Jahii Carson stuff (Jahishalls are real). Stay tuned and thanks for reading through this 1567 word marathon.