Investigating the effectiveness of press coverage

By Matt Edwards | June 14, 2024 | 1 min read

Investigating the effectiveness of press coverage

“To press, or not to press, that is the question. Whether ‘tis nobler in the mind to suffer the easy catches, and yards playing off leverage,or to take up arms against a sea of receivers, and by pressing disrupt them?”

What’s that? That’s not how that goes? That’s too bad. I can only imagine how much more fun it would have been to read Shakespeare in school if he was talking about defensive back play.

With Wide Receivers getting more and more talented (don’t believe me, just look up 6’5” 240 LB South Carolina WR Nyck Harbor and his track times), rule changes that benefit passing games, and a general shift in schematics, a Corner’s job is getting more and more difficult. One of the last vestiges of physical play at the position comes in the battle at the line of scrimmage.

Press Coverage Defined

First, let’s talk about press coverage, and how we at StatsBomb define it. One of the great things about StatsBomb data is the ability to build rules and models to “tag” plays instantly without need for manual intervention. Using the locations at the snap of all-22 players we are able to quickly identify individual skill players, formations, motions, defensive structures, and much more. One of the things we use the snap location for is identifying defensive backs who are pressing receivers. Using a threshold of 2 yards from the line of scrimmage, and 1 yard to either side of a receivers location, we are able to identify every time a player is lined up in press coverage.

To get a feel for what that looks like in practice, below are the snap location charts for the corners who pressed the most and least in the NFL and FBS.

->
->

Tyrek Funderburk and Kendall Fuller were the least pressing corners in FBS and the NFL respectively, while AJ Terrll and MJ Devonshire led their leagues in press %. Press coverage in and of itself is a binary yes or no, but that doesn’t mean that all “off” leverages are equal. For example, although Funderburk and Fuller were both limited in pressing, Fuller mixed up his alignment depth much more than Funderburk.

To Press or not to Press?

There are many reasons why a team might try to press receivers, but what does the data say about it? Now that we have defined what press coverage is, and taken a look at some extreme examples of how it is deployed, let’s answer the question that Shakespeare (ahem, I) coined at the beginning of the article: to press, or not to press.

1. Get closer to a WR

The biggest factor in both our Completion Percentage Over Expected (CPOE) and Catch Rate Over Expected (CROE) models is the separation of the receiver from the closest defender.

If a defense wants to decrease the likelihood that a pass is completed, the number one thing they can do is have defenders closer to the receivers. That seems intuitive when you think about it, it’s easier for a receiver to catch the ball if he doesn’t have anyone defending against him.

But how can a defense get defenders closer to the receivers? Well, one of the easiest ways is to start closer to the receivers. The plots below show all qualifying FBS and NFL corners and their separation at the time of target. As you can see, the vast majority of players had lower separation while pressing than when they played off coverage.

->
->

Each of the top 7 corners taken in this year's NFL Draft had lower separation while pressing. Cooper DeJean had the biggest difference allowing 2 yards of separation while pressing, to over 4.5 yards while not pressing. If the goal is to have defenders closer to target receivers, then the answer is a resounding Press.

2. Allow Fewer Targets

Paraphrasing the great quote from Michael Scott (who is paraphrasing from Wayne Gretzky) a DB allows 0% completions on passes that don’t get thrown to the receiver they are covering.

Lots of things go into when a receiver is targeted with a pass: pressure on the QB, where the receiver is in the schematic progression, and of course how well the receiver is covered. We just examined press coverage and the relationship it has with separation. Next, let’s look at press coverage and the relationship between target percentage.

->
->

Defenders at both levels are more likely to allow a target while pressing than when they aren’t. As far as target % goes, the data shows Not to Press.

3. Success when targeting receivers

The next level of analysis is when the receiver is actually targeted. Does press coverage have any bearing on success of the pass attempt? There are many ways we could examine the success of a pass. Completion percentage, Catch Rate Over Expected, and more. For this example, and many examples, EPA is a good estimate. Our Catch Rate Over Expected model already take into the separation which we already looked at. EPA takes into account down, distance, field position, and is a metric that gives more context to the situation than just binary complete or not.

->
->

Instead of a linear representation line on these plots, there is a trend line for the data. And that trend line is about as flat as can be, if the goal is to have more unsuccessful pass attempts when the receiver is targeted, it does not matter.

Introducing a new stat!

One of my favorite things to do in my work is exploring things that I haven’t seen elsewhere before. Occasionally this will weasel its way into a more drawn out thought process, make it into our data somewhere, and maybe even get an article written about it (see RANDY).

The first three areas examined above have left us in a deadlock.

  1. Being closer to the receiver - Press
  2. Allow fewer targets - Not Press
  3. EPA allowed - Doesn’t Matter

This leads us to our last and final category: downfield movement. We have spent a lot of time at StatsBomb recently looking into our tracking data and physical metrics. Tracking Data gives us the speed and acceleration for every player 30 times per second. Another thing we collect with the physical metrics is the distance traveled between frames. These roll up into play level metrics like top speed, top acceleration, total distance. We then contextualize these physical metrics using our eventing data to create football specific metrics like defender get off distance, db closing speed, receiver deep route speed, and more.

The first step when creating a defensive back press coverage metric was defining the parameters. Obviously creating the matchup with the defenders that are in press coverage and the receivers. Only include plays where the time to throw is greater than or equal to 2 seconds to eliminate screens or other quick throws. With those in place, the actual measurement comes from how far a receiver gets downfield in 1.5 seconds.

In college, the average distance a receiver gets in a non-pressed situation is 3.36, while the average distance in a pressed situation is 2.65 yards. In the NFL, those numbers follow a similar pattern. The distance in non-pressed situations is 3.14, and the average distance while pressed is 2.63. To round out the overall analysis, in the case of downfield distance, PRESS.

For some quick reference, the top qualifying corner in FBS last season was Dorian Strong out of Virginia Tech. His average downfield distance allowed was 1.94 yards. We’ve looked at some of the top corners in the draft in plots above, and the top two of that group were Ennis Rakestraw and Kool-Aid McKinstry in 14th and 15th nationally. In the NFL, Joshua Williams of the Kansas City Chiefs was number 1 with a distance allowed of 1.8 yards. All Pro Corner DaRon Bland was last (among qualifying Corners) with 3.65 yards allowed.

Conclusion

There are many reasons a defense might decide to press a receiver. We discussed a few in this article: trying to limit receiver separation, allowing fewer targets, allowing less success when the pass is attempted, and limiting downfield movement. And this didn’t take into account down and distance, scheme, individual matchups, and much more that a coach might think about while putting together a scheme. I’m not sure what the answer was for Hamlet (Apologies to my AP English teacher Mrs Gregg); but in terms of the football version of Hamlet’s soliloquy, there is not a definitive answer.

Matt Edwards
Head of American Football Analysis - StatsBomb
matt.edwards@statsbomb.com
@thecoachedwards on X (Formerly Twitter)

By Matt Edwards | June 14, 2024