Nylon Calculus: What dribbling does for NBA shooters

If you give a ball to anybody on a court, you’ll see them dribble a few times before taking a shot. It’s a natural setup for shooting a basketball, like taking a practice swing in baseball, and even the best free throw routines usually involve a dribble or two.

Yet among basketball fans and in the analytics community, there are special words for players who dribble too much before a shot: “ball stopper”, “not in the flow” and so on. The underlying belief is that a high-dribble shot is often bad for offensive efficiency.

Today, I’ll take a look at who are the heavy dribblers, the effective dribblers and if there is an optimal number of dribbles before a shot. Dribble stats are broken down by dribbles before a field goal attempt, which can be found on NBA.com. I cleaned the data for low-frequency players (minimum 20 games played and 100 field goal attempts) and low-dribbling players (with the assumption that heavily zero dribble FGA players are lob men, dunkers or catch and shooters). This leaves me with a sample of players who can and do dribble before they shoot.

Read More: Missed 3-pointers are ruining Kawhi Leonard’s defensive numbers

Next, I calculated dribbles per shot by applying a number of dribbles per shot attempt (I used 4.5 dribbles for 3-6 dribbles FGAs and eight dribbles for 7+ dribbles FGAs). I then did some grouping, summing and averaging to get to dribbles per field goal attempt or DrFGA. Below is DrFGA by eligible player across the league:

Within our group of players, DrFGA is a steady slope which should mean that we have a decent representation of players who dribble and don’t dribble. Let’s take a peek at who make up the far left and right of that chart.

LOWEST DrFGA
Marvin Williams 0.496
Jonas Jerebko 0.513
Brandon Rush 0.578
Alex Abrines 0.621
Sam Dekker 0.623

HIGHEST DrFGA
Jeff Teague 4.717
Reggie Jackson 4.892
John Wall 4.969
Kay Felder 5.073
Tyler Ulis 5.143

In the low dribble group, we see guys who aren’t known for their playmaking and lean on the taller side. In the high-dribble group, we see a lot of ball-dominant point guards who run their team and bring the ball up on offense, so this makes intuitive sense. Next, let’s get to efficiency to confirm the “less-dribbles” theory and we’ll break this down into 2-pointers and 3-pointers. First, 2-point shots:

In this 2-pointer field goal percentage chart, there’s a clear the advantage to not dribbling or taking only one dribble (probably assisted field goals or dunks) but there isn’t really a disadvantage to dribbling more once you’re at two dribbles either. The 3-point field goal chart shows this more clearly:

At zero dribbles, the performance range is quite narrow. These are your catch-and-shoots – generally open shots. At one dribble, the efficiency peak flattens dramatically. Beyond a single dribble, the shapes look quite uniform. This suggests that once you take two dribbles, it’s all about the same again. Despite the spread in performances, the middle point of all the shapes are relatively clustered and overlapping — it’s not as dramatic as you might think.

Let’s see how this looks in a scatterplot. We’ll focus on 3-point percentage since that does a better job of taking away easy shots like dunks or transition buckets.

This plot suggests dribbling has a positive effect on 3-point percentage. It shows that an extra dribble is worth 500 bps on field goal percentage, meaning four dribbles is worth around a two percentage point lift to 3-point percentage. The OLS regression shows a p-value of .1 so there is some chance this is chance-driven. We also do see that the number of players go down as Dr3FGA goes up, which may suggest that only the best players get to be heavy dribblers.

Let’s see if the best players are dribbling their way into the high Dr3FGA brackets by partitioning the data into the high Dr3FGA players and the rest of the league below the 75th percentile as a control.

What’s interesting is the gap in 3-point percentage performance at 7+ dribbles. In fact, the 7+ dribble 3-pointer by the high DrFGA player is the second best shot in that chart. High Dr3FGA players hit their 7+ dribbles shots (37.5 percent) almost at the rate of a zero dribble shot, while low DrFGA make only 30 percent of their 3-pointers.

Lower Dr3FGA shooters get progressively worse and worse as they dribble, which is probably what creates the perception of heavy dribbling hurting offenses. They do see an uptick at 7+ dribbles but this is a much worse shot than a single-dribble shot for them. Ironically, their one-dribble shot is better than a one-dribble shot by a heavy dribbler.

What this means is that the theory that dribbling hurts shooting efficiency is that the theory doesn’t apply uniformly. Players that aren’t heavy dribblers get into trouble when they start dribbling (excluding the uptick at 7+ dribbles) while players that dribble a lot, might get a better shot off if they dribble some more.

Next: Nylon Calculus -- The curious case of Boogie and the Brow

Obviously, there are other factors to consider like energy exerted by shooter, team stagnation but in isolation, a heavy dribbler dribbling it out appears to shoot better than forcing a quick shot to be in flow. The next time you see a Kyrie Irving, Chris Paul or Goran Dragic dribbling out a 3-pointer, know that it might be almost like hitting a wide-open 3-pointer.