By: Richard Li
[caption id="" align="aligncenter" width="300"]
Click for a full-sized interactive version.[/caption]
Last week was interesting. I heard so much about these balls being stopped and those balls being sticky that I thought I was writing for an adult website. Dirty jokes aside, the basketball version of balls that are stopped, and which may or may not be sticky, is a topic worth exploring because most people generally agree that stopped balls are not good. The hard part is actually identifying data that captures who makes balls stick. Just in our forums alone people batted around ideas and exceptions to ideas. What I do in this post is create a compilation of different measures that collectively identify which players make balls stop the most.
Measures
Measure 1: Usage rate. I feel like usage rate is a somewhat misunderstood statistic, maybe just due to its name. A more accurate and longer description is how many possessions end with a certain player while that player is on the floor. Ending a possession means a shot, foul shots, or a turnover. So if player X has a 25% usage rate, that means 25% of all possessions end with player X while he is playing. The disadvantage of usage rate for our purposes is that it doesn't capture everything else that happens before the possession ends. Also, a turnover might be a result of a ball moving play, like if the player was trying to pass the ball.
Measure 2: Touches per pass. I developed this for my selfishness posts. It's exactly what it sounds like, number of touches divided by number of passes. It's better than usage rate at capturing what happens before a possession ends, since its unit of analysis is each touch instead of each possession. What it does not do, however, is capture what player X is doing while he's touching it, or for how long he touches it.
Measure 3: Seconds per touch. Henry Abbott was really interested in how long players hold the ball while they're touching it. Passing it immediately, he said, is different from holding it, jab stepping, and letting the play die. So I developed this statistic to measure how long someone holds the ball every time they touch it. It's simple, just seconds of possession per game divided by touches per game.
Measure 4: Outside touches %. In the forums, someone also mentioned that touching and holding the ball close to the basket is different than touching and holding it far from the basket. After all, if you're five feet away from the basket, you should touch the ball because you have a better chance of scoring. I created this statistic to measure what percentage of a player's touches originate far away from the basket, defined as not in the paint and not at the elbow.
The Table
The table above shows values for all players who played at least 35 games and at least 22 minutes per game. Each player is also ranked in each statistic. The final column is a rank of ranks. Basically, each of the players ranked is added and that number is then ranked in this column. Assuming that each statistic has equal "ball stopping power," then whoever has the highest overall rank stops the ball the most.
Note that the data for touches per pass are from my last selfishness post, for which I collected data in late January. Moreover, due to a few players who were in the touches per pass data and not in the more recent data, or vice versa, there may be some slight inconsistencies in the ranks. Some players who switched teams also have very small sample sizes. All these cases were removed.
Observations
I despise that NBA.com's data does not identify player position. It's especially crappy in analysis such as this one, in which one position (point guards) are heavily biased. As one would expect, point guards have and hold onto the ball the most, and tend to start far away from the basket, like in the back court. Not much they can do about that. So we can compare point guards to each other, but it's hard to compare other positions to them.
No surprise that James Harden is the most ball stopping Rocket. It is a little surprising just how high he ranks compared to everyone else, including point guards. In fact, he's fourth overall, ahead of players like John Wall and Brandon Jennings. Maybe Henry Abbott (and Chandler Parsons) was onto something.
Discuss.
. 
