clock menu more-arrow no yes mobile

Filed under:

Do 'impact plays' truly impact the defense?

Do highlight plays make a defense, and do the Chargers need more of them?

Jahleel Addae & Melvin Ingram attempt to bring down a sneezing Ryan Tannehill
Jahleel Addae & Melvin Ingram attempt to bring down a sneezing Ryan Tannehill
Sean M. Haffey/Getty Images

We live in a sports world inundated with highlights, and we often lose sight of all the other snaps that can make just as much of an impact when put together. That said, can we really discount them? Are they overrated or are they the backbone of a top caliber defense? (Or, like many things, does the answer lie somewhere in between?). Is it possible to be a top defense without being "flashy"?

In this piece, we will be looking to see if there is a correlation/connection between frequency of "big plays" & overall quality of defense. For this exercise, we will consider all of the following as "big plays": QB sacks, run-stuffs (run-tackles for gains of 0 or less), INTs, passes defensed (not including INTs), and forced fumbles. There are variables in the quality of these stats (ie: barely grazed PDs or FFs that bounced right to another O-player or occurred as the same play as a sack or stuff), but for simplicity’s sake, we’re going to group them all together.

Before I reveal the entire data table, let’s take a look at some plots of said data (as the color coding in the table may "spoil" trends too soon).

This first scatter plot uses Impact-Plays-per-Game to measure  highlight-rate and Points-per-Game to measure overall defense quality.


IP/G is along the horizontal axis and PPG-allowed along the vertical. The Chargers’ point is indicated by the red arrow. While spread out, we can notice there is still a downward trend (as IP/G goes up, PPG allowed goes down) – indicated by a line of best fit. In the top right corner, you will find the equation for the line of best fit and the r-squared value. To briefly explain, the r-squared value shows how closely the data fits the line, where 1.0 would be the perfect fit (ex: wins vs. loses (assuming ties not possible) would result in a 1.0, as any NFL team with 10 wins would have to have 6 losses, 6W-10L, etc). Our r-squared value in this chart is 0.23, which while not great, is not terrible – especially since an NFL game represents a system with many variables at play (many of which are difficult to empirically measure). However, I know we can find better ways to measure our variables…

Let’s now plot IP percent against Football Outsider’s defensive DVOA% (where a negative value is better since you are holding the opponent further below the average per-play value).


You can notice that the spread here is much tighter. This is reflected in the considerably larger r-squared value. The Chargers’ point lies above the line, meaning that their DVOA was lower than expected given their IP%. Looking at this chart you may notice there are 2 clear visual-outliers. These are the entries for Denver (1st in DVOA) & New Orleans (32nd in DVOA), who were considerably better & worse (respectively) than the line would have predicted. If we were to remove these 2 entries from the sample, we get an ever larger r-squared value – roughly 0.74.

Finally, here is the complete data table, sorted by DVOA, and for convenience, a chart with the Chargers’ rankings in each IP category

IP Data Table

With the color-coding, we can notice the strong correlation between DVOA & IP%. We can also notice that the Chargers don’t rank well in any category. They rank near the top of the bottom third in most categories, and a dismal 28th in DVOA (the final result). As I noted earlier, the Chargers’ DVOA is worse than we’d expect, even with their low IP%. While not included in these charts, the Chargers’ mark of 5.8 worse DVOA than expected is also 28th worst. They are not making many plays, and they are not playing very well on all the plays in between. (You can read this fan post by arandom for a look into what may be ailing them on these plays).

So, do big plays matter? Should the Chargers be making more of them – or rather, would making more of them (because obviously you want more big plays) really help them? Well, no team in the top 10 of IP% finished lower than 12th in DVOA. Good DVOA defenses are making a lot of impact plays. Do you want to be a good DVOA defense?  Well, the top-10 features 8 playoff teams (with the Jets falling just short, due to the erratic play of QB Ryan Fitzpatrick). The Chargers need more impact plays, bottom line.

I think with the addition of Joey Bosa (& the benefit Corey Liuget receives), Melvin Ingram’s late-season emergence, and – fingers crossed – a healthy Jason Verrett, I’d wager there’s a strong chance we see more impact plays & thus, a higher DVOA ranking in 2016…fingers crossed.