Analytics, Basic Stats, and Recent Historic Context: The ABC’s of Penn State Football – The 2020 Defense Retrospective

Introduction

Now, it’s the defense’s turn. Same general premise as yesterday. As we turn the page to 2021, let’s take one last look back on what was the 2020 Penn State Nittany Lions – specifically defensive coordinator Brent Pry’s unit. We’ll explore what trended well, and explore what needs to be fixed in the offseason. And how are we going to do that, you ask? Same way we did weekly during the season – using a mix of basic and advanced stats within the broader context of the James Franklin era at PSU. 

Basic Statistics Summary

You’ve watched all the games but it’s worth reposting the basic stats summaries. Let’s start with yards allowed. In the graph below we have the passing yards (blue), rush yards (orange), and total yards (red) surrendered in each game.

Two things immediately jump out to me here:

  • How in the world did Penn State lose the Indiana game? The Nittany Lions defense allowed a meager 211 total yards against the Hoosiers…and lost? Throughout the 88 games James Franklin has coached from the Penn State sidelines, his teams have held opponents to 211 yards or less 10 times. Franklin’s record in those 10 games? 9-1. Opponents in those 10 games include Rutgers (3X), Maryland (2X), Idaho, and Akron. No one should lose games where the opposing offense generates 211 yards AND where your offense has 488 yards. But we know what happened. The offense had too many turnovers in IU territory, couldn’t convert red zone
  • The lower water mark was the Ohio State game where the defense allowed 526 yards. That’s the 5th most yards allowed in the Coach Franklin Era, trailing only: USC–Rose Bowl (575 yards), 2018 Indiana (554), Memphis–Cotton Bowl (542), 2017 Ohio State (529). The witness-less Whiteout marked only the sixth time a Franklin team allowed more than 500 yards (the other being 2016 Michigan). Penn State’s record in these games is 2-4.

Since 2014, Penn State’s defenses have given up an average of 130 rush yards per game, 201 passing yards per game, and 331 total yards. The 2020 Lions allowed 129, 198, and 327 in the respective categories which is nearly SPOT-ON average for a Penn State defense. 

Moving to scoring defense, we see a dramatic difference between wins and losses. In the plot below, we have touchdowns allowed (orange) and total points allowed (blue) by game. In the first five games, the team (including the offense on pick-6’s or scoop-and-scores) allowed at least 30 points and 3 touchdowns. In the four wins, Penn State allowed no more than 24 points. 

When you allow 30+ points per game, you’re going to lose a lot. I mean, this is the Big Ten, not the Big 12.  Since 2014, Penn State has allowed 30 or more points in 22 games. Its record in those games is 5-17 (.227 win percentage). Conversely, when the Lions hold foes to 24 points or fewer, which has happened 54 times since 2014, their record is 47-7 (.870 win percentage). For those wondering, 12 times games under CJF Penn State gave up between 25 to 29 points. Penn State is 5-7 in those games. 

To summarize, don’t give up 500 yards and/or 30+ points. #Analysis.

Advanced Stats

What, if anything, can we point to within the advanced stats that contributed to that reduction in points allowed post-Week 5…besides fewer turnovers in opponents’ territory by the offense, of course? Let’s explore the categories that matter: Red Zone Efficiency, Havoc Created (turnovers, sacks, TFL), Explosiveness, etc. 

Possession Efficiency

Possession Efficiency is one of my new favorite stats, especially when viewed in the context of points and yards. Here is the game-by-game performance of Penn State’s defense in yards-per-possession (x-axis) and points-per-possession (y-axis). Since 2016, Big Ten defenses average 30 yards-allowed-per-possession and 2.12 points-allowed-per-possession. 

For those wondering, the number below each opponent’s logo is the total points score vs. PSU. The straight, evenly-rising line represents the “expected” results as yards and points increase. Any data point above or below the two curved lines would be an “outlier” result – for instance, with its Yardage Efficiency Rutgers SHOULD have scored more (although not much more) than 7 points. Conversely, Indiana and Iowa SHOULD have scored fewer points based on their Yardage Efficiency.

When a team like Ohio State generates 48 yards-per-possession, it will typically lead to more points (>3.4 PPP and 38 total points). This is expected. And, on the other hand, if a defense limits the yards allowed, they’ll often limit points, as well. Let’s compare these individual 2020 games to the entire Big Ten’s performance from 2016-2020 in the graph below.

In the Big Ten graph, you can see that there’s a fairly linear trend. If a defense is good enough to surrender 24 or fewer yards-per-possession, it’s expected to hold opponents to less than 1.4 points-per-possession (with a couple of outliers). 

With that established, let’s contextualize the 2020 games against Indiana and Iowa. Indiana scored 2.6 points-per-possession even though the Hoosiers only gained 15 yards-per-possession; Iowa scored more 2.9 PPP on 26 YPP. Those are INCREDIBLE outliers compared to the seasonal averages. To register 2.6 points-per-possession, Indiana should’ve needed to gain 36 yards-per-possession. Iowa’s 2.9 PPP output would’ve been expected from 38+ YPP, not a measly 26 YPP. Those two games highlight the struggles (and a bit of bad luck) of Penn State’s defense early in the season. Again, the fault is not solely on the defense but a high contributor is the bad positions they were put in by the offense with turnovers and general early-season ineptitude.

Red Zone Defense

“Performance in the Red Zone matters” seems like something I’ve heard a lot while watching football. For the 2020 Penn State football team, it mattered but also wasn’t very black/white either. Below is the number of RZ attempts allowed (x-axis), the points allowed per RZ attempt (y-axis), and the total opponent red zone points (label). 

In four of the five losses, opponents reached the red zone at least four times and scored more than 16 RZ points with a 4-point average per RZ attempt. In the 5th loss, Maryland reached the red zone exactly zero times but still scored 35 points. In three of the 4 wins, opponents converted every red zone opportunity to a touchdown (Rutgers 1 TD, Illinois and Michigan 2 TD each). The one “trend” that can be inferred is that in the last four games, the Penn State defense only allowed 2 RZ attempts per game versus an average of 4 in the 5 losses. This is an improvement in overall production albeit a reduction in efficiency when it came to preventing red zone scores. 

Havoc Generation

Havoc is the sum of fumbles, interceptions, tackles-for-loss, and sacks that a defense generates as a percentage of total offensive plays against (note: This stat was originally developed by Bill Connelly at ESPN and his stat includes passes broken up, whereas I do not include it.). It is an indicator of a defense’s aggressiveness and overall goodness. An average Big Ten team will have a 14% Havoc Rate. Between 2016 and 2019, Penn State’s average Havoc Rate in conference games was 16.5%  — so, above average. The 2020 installation of the PSU defense had an inconsistent performance in Havoc Rate, as shown below. They bounced between good and bad for the first seven games of the season and closed with two exceptional performances against Michigan State and Illinois with 20+% Havoc in each game. 

Some of these percentages you’d expect. For instance, a 9% Havoc Rate vs. Ohio State – with its strong offensive line and future multi-millionaire QB – doesn’t seem out of place. In that game, Penn State produced 6 Havoc plays. But against Iowa, who similarly has a good offensive line but a less-talented quarterback, Penn State managed 19% Havoc…and still lost by 21 points! Then, in their first win, against Michigan, the defense only generated 4 Havoc plays. Typically, Havoc Rate is a statistic that has a relatively good correlation to points allowed and overall defensive efficiency. This year though, not so much. However, the upward trend post-Michigan through Illinois is positive and may be something to build on going forward.

Explosivity

Defensive Explosivity (i.e. limiting explosive plays by the opposing offense) is calculated by: Yards x Points/Plays/100. Regular readers will know that Penn State’s offense has been poor in this regard in 2020. The defensive performance is shown below — the x-axis is just the points-allowed-per-play and the y-axis is the explosivity calculation described above. It’s interesting to look at these two together. Since 2016, Big Ten teams score an average of 0.37 points/play with an explosivity rating of 1.46.

Compare that 0.37 to the x-axis of the five losses. In each of those defeats, the PSU defense gave up at least 0.48 points-per-play. In every loss, except Indiana, the opposing offense qualified as “explosive” vs. PSU’s defense. This is analogous to the Possession Efficiency described earlier. Indiana scored at an incredibly high rate in points-per-play metric and points-per-possession but was below average in explosiveness and moving the ball. Maryland was slightly above average in efficiency but was incredibly explosive (4 offensive touchdowns on drives of 4, 6, 6, and 8 plays and a defensive score; they also had 6 drives that averaged 11 yards). In Penn State’s four wins, Pry’s group limited explosiveness and didn’t give up big plays OR defensive scores (which is an implicit factor in the explosiveness metric and one that I like keeping in since a team scores 7 points on zero plays). 

Overall Defensive Effectiveness

When we put it all together we get “Defensive Effectiveness” (Def-Eff) as a single number to show how good a unit is/was. For defense, lower is better and when we plot game-by-game we see a strong demarcation between the five losses and four wins. An average Big Ten defense has a Def-Eff of 26.7. In Penn State’s losses, they were never below 31.6 (Indiana) with bad performances against OSU and Maryland. In their wins, they were always better than average. 

For the full season, Penn State’s Def-Eff was 27.6 which is easily the worst defense vs. conference opponents in the past five years: 2016 (20.8), 2017 (18.5), 2018 (19.1), 2019 (14.9). But, averaging the last four wins alone gives PSU a 16.9 Def-Eff value, which would rank as the 2nd best Penn State team since 2016. Granted, the four offenses Penn State played to end the season weren’t exactly the Greatest Show on Turf, but I’d much rather go into an offseason with the defense’s momentum trending positive rather than the other way around. 

 

Pattern photo created by jcomp – www.freepik.com