# Nylon Calculus: An NBA block is like an NFL sack

Sacks were once the glamor stat on defense in the NFL. Interceptions and fumbles happened too infrequently, tackles too frequently and so sacks became the gold standard for measuring front line defensive players. The comical irony of this culminated in Michael Strahan’s “sack” of Brett Favre to break the NFL single season sack record.

Still, in the era of Moneyball, football fans started to recognize the sack as a lagging indicator of quarterback pressures. Around the same time, the correlation of coverage time to sacks was linked and today, most people agree that sacks is a pretty narrow player evaluation tool.

For the NBA, the block is the most celebrated defensive statistic (probably because it’s easier to package as a SportsCenter highlight than a steal). It also happens to be the most visible stat that can be easily traced back to player height, which the NBA specializes in.

However, like the sack’s relationship to quarterback pressure, it could be asked “are blocks representative of rim protection?”. We often see players who time blocks perfectly but also shirk from the basket for the fear of ending up on a poster. And like cornerbacks and safeties do for sacks, do perimeter defenders in basketball have a role in rim protection stats?

I will try to answer these questions today. The players I’m analyzing have played in 20 or more games and average more than 10 minutes per game as of March 16. Data is courtesy of NBA.com.

### Where to find defensive rim protection cheaply

We’ll start with a quick definition of defensive rim protection from NBA.com:

Defensive Impact: Statistics measuring the impact a player has on defense, including blocks, steals and protecting the rim, which measures the opponent’s field goal percentage at the rim while it is being defended. Rim protection is defined as the defender being within five feet of the basket and within five feet of the offensive player attempting the shot.

The following chart looks at the defensive rim field goal percentage by players across the league with a definition for defensive impact below it:

Those tails look super interesting – let’s look at the top ten and bottom ten players:

That’s a surprising number of guards and forwards at the extreme ends of defensive rim field goal percentage. Non-centers get less defensive rim field goal attempts, which reduces their value, but it appears that those players have a wider range in defensive performance. Below is a plot of defensive field goal attempts per game against defensive rim field goal percentage:

This is actually a positive skew (left-leaning) distribution when you account for the fact that the base of the population is more in the .45 to .7 range. The top of the distribution skews left so we know, generally, that defenders who contest a lot of shots at the rim do better than those who don’t but what’s interesting is the wide range of field goal percentage for the low volume rim protectors.

To take an example, compare two point guards from the earlier table, Jeremy Lin and Brandon Knight. Lin has a .432 defensive rim field goal percentage while Knight has a .678. The ensuing difference on Lin’s two defended rim field goal attempts per game difference is .492 made shots, basically a point every game, which doesn’t sound like a lot but it is.

Rudy Gobert, who is also on the same table, has a .433 defensive rim field goal percentage on 10.3 defended field goal attempts, the highest in the league. If you add half a made shot per game to Gobert’s numbers, his defensive rim field goal percentage becomes .500, which is not far from Mason Plumlee. If you played Plumlee in place of Gobert, that would represent an enormous change in rim protection/attacking values yet do teams account for this when Lin is on the court vs. a low-end defensive rim guarding guard?

Due to Gobert’s high number of contested shots at the rim, an additional half-field goal made is actually dampened. Applying the same half-field goal contorts DeAndre Jordan (495 DefRim DFGA at 7.8 Def Rim FGAs) to look like Nikola Vucevic or Gorgei Dieng (both at .529 DefRim DFGA), essentially replacing a 75th percentile rim defender with a 40th percentile rim defender.

Now Lin’s particular numbers may be a product of small sample size and across the range of low volume rim protectors, we probably see a lot of noise. Still, the implication is important. A team can probably improve their paint defense if they can identify perimeter defenders who are impactful defending the rim, even at low volumes. Those players could probably be found much more cheaply than bidding for the premium interior defended. Obviously, there are tradeoffs to be made but for a team looking to improve rim protection, perimeter players provides a source of value.

### Do good small rim defenders and good big rim defenders go together?

In our NFL analogy, good cornerbacks make defensive ends look better. Is there a similar relationship between shorter perimeter players (low volume rim protection) and taller players (high volume rim protection) in the NBA?

I partitioned players into shorter and taller players (using 6-foot-10 as the cutoff) and recalculated defensive rim field goal percentage by team. This will miscategorize an occasional short rim protecting specialist like Draymond Green but should serve most of the NBA.

The correlation between big man defensive rim field goal percentage vs small defensive rim field goal percentage by team is basically non-existent (correlation coefficient of .164). This implies that having superior interior rim protectors and perimeters rim protectors don’t necessarily go hand in hand. However this data is just from this season and doesn’t necessarily disqualify the potential advantage if a team were intentionally building their roster to exploit it.

### Blocks are like sacks

So do blocks predict defensive rim field goal percentage? Or is some defensive other stat more predictive? After pulling in a several more numbers from the NBA’s hustle and general defensive stats which covers charges drawn, steals, deflections, defensive rebounds, points off turnovers and more, I ran a Principal Components Analysis. What came back as the two primary principal components for predicting defensive rim field goal percentage were blocks ,but also second chance points surrendered. In fact, second chance points were a bigger component than blocks (39 percent vs 29 percent) for explaining defensive rim field goal percentage variance. The intuition behind this is probably that second choice points happen close to the rim so it’s actually more important to box out than it is to get the swat at the rim.

Within the variables I used (which doesn’t get into player tracking, lineups, defender distance, etc), 70 percent of the variance in defensive rim field goal percentage can be explained in those two variables, which is decent. In summary, blocks are predictive of defensive rim field goal percentage as you might expect but they don’t nearly explain the full picture.

In summary — don’t undervalue perimeter defenders when it comes to rim protection because there is more value to be found. Although one year of data seems to indicate that rim protection is more of an individual thing — better smaller defenders and big defenders don’t necessarily or augment pair together. As to blocks, they are a highlight stat but also are fairly predictive of defensive rim field goal percentage. However, securing key defensive rebounds are just as important to interior defense.