Is There a Strong Relationship Between Defender Distance and Field Goal Percentage in the NBA?
With more and more of STATS' SportVU data being made available to the public, we are now able to discern information that we wouldnâ€™t have even dreamed of several years ago. Among the interesting new data SportVU provides is player distance on the basketball court. These amazing cameras are able to capture where each player is located at all times.
Obviously, this is groundbreaking in evaluating spacing within both offenses and defenses. Spacing is a trendy word right now in basketball â€“- offenses want to create it and defenses want to limit it. The stretch big movement is a direction correlated to this, as the more shooting you have at each position, the more you can space out your offense.
Another thing we can look at now is how much space is optimal for jump shooters. SportVU tracks â€œcontestedâ€ and â€œuncontestedâ€ shots, with the line being four feet of space. So, if a player has a defender within four feet of him, the shot is considered contested. If not, itâ€™s considered an open shot.
Creating a strict line like this is useful, but it doesn't tell the whole story. Four feet is a lot of space â€“- is there a discernible difference in shooting efficiency between, say, a defender being four feet away and a defender being just one foot away? You would think so, right?
Thanks to smart computer guys and cool sites like that are on the forefront of the analytics movement - like numberFire! - we can actually look at shooting efficiency at every single point of defender distance. So letâ€™s do just that.
It might seem odd that field goal percentage actually goes up the closer a defender is to him, but it actually isnâ€™t. The graph (which Iâ€™ve made interactive here if you want to see the FG% at every distance point) includes every shot of the 2013-2014 NBA season, so obviously dunks and layups are included in this data set. Naturally, the most player congestion is around the rim, which is also the place where shots are the most efficient, so we have a large number of shots that are â€œcontested,â€ but still highly efficient.
For example, if Chris Paul drives into the lane and dishes to Blake Griffin, who slams it through, that's obviously a highly-efficient shot. However, if there is any defender within four feet, regardless of what theyâ€™re doing â€“ they could actually be contesting the dunk, or rotating too late, or letting him have it to avoid foul trouble, or just lying on the ground â€“ that shot is deemed â€œcontested.â€ As you can see, itâ€™s a loose definition that can muddy up the findings.
Since this fact really skews contested versus uncontested shooting data, we should probably narrow the parameters. What does field goal percentage look like at each defender distance point for shots eight feet or more? Looking at shots outside the paint and restricted area focuses on contested versus uncontested jump shots, which is more what weâ€™re looking for.
This graph shows more of what weâ€™d expect, although itâ€™s still far from perfectly linear. As you can see in another interactive graph, field goal percentage with defenders 0.9 feet away was 27.4%, but was 38.7% when defenders were 0.6 feet away. The reason for this is probably small sample size -â€“ there were tons of shots to analyze between 1-4 feet, but it got less as it approached zero. If we expanded the data to several years, this would probably normalize.
Before this project, I didnâ€™t entirely understand the four-foot line. If a defender is there or farther, itâ€™s uncontested, but if closer itâ€™s contested, even if itâ€™s 3.9. I thought it was more-or-less an arbitrary number, but it seems like there is some merit to it. Right at that four-foot defender distance mark is where we see a noticeable dip and continual trend down in regards to field goal percentage. Perhaps four feet really is the correct line.
The graph also shows that around the 1.5-foot mark is where shooters really struggle on a consistent basis. From four feet to that mark, itâ€™s a steady decrease, but itâ€™s not a steep drop. However, between 1.5 and 0 feet of defender distance (assuming random points would normalize over years), field goal percentage is very low, which is pretty intuitive.
Now, letâ€™s make this applicable to this current season. Which teams so far are maximizing these types of shots (greater than eight feet and a defender within, say, two feet [for simplicityâ€™s sake])? Here you go. Also, itâ€™s interactive again here.
If you want to find a reason why the Portland Trail Blazers have been very good defensively this year and want to look beyond the obvious Robin-Lopez-and-Chris-Kaman rim protecting thing, here's your answer. They've been highly contesting their opponentsâ€™ jump shots. Those shots against the Blazers â€“ and the Spurs, but thatâ€™s no surprise â€“ have had a defender within that two-feet zone over 11% of the time.
Also, we would expect the Bulls and Grizzlies to be at the top of this list thanks to great defenders in Jimmy Butler, Mike Conley and Tony Allen, and indeed that is true. The Utah Jazz, currently boasting the leagueâ€™s second-worst defense, is the odd one at the top of the group. However, it could be a sign of good things to come for the Jazz â€“- theyâ€™ve faced a brutal schedule lately and are still a young team in the Western Conference. Though, if they can continue to defend on the perimeter like this and can continue to develop rim protection with Derrick Favors and Rudy Gobert, they could be good defensively in the future.
Nonetheless, weâ€™re still looking at a small sample for this young season, so perhaps teams like the Jazz and Pistons regress back to where weâ€™d expect them with such bad defenses. Next time you watch a game, pay attention to how close your team defender is to shooters. It might give you some clues about how your defense will perform in the future.