And now, for the main event. We’re here to perform an autopsy on the 2021 Georgia Tech defense.
Here’s the conclusion we shared in this same article from last offseason:
For the Georgia Tech defense in 2021, disruption up front and consistency in pass coverage will tell the story of the season.
They certainly did tell the story of the season. Just not at all in the way that we hoped.
The defense disappointed in 2020, and then defied all expectation of even the most doomsday fan this past season. Look at the change in some of the key defensive metrics from 2020 to 2021:
Opponent Success Rate: 45% —> 48%
Opponent Yards per Dropback: 7.2 —> 8.9
Opponent EPA/Pass: 0.22 —> 0.41
Opponent CPOE: -0.8% —> 8.4%
GT Pressure Rate: 23% —> 22%
GT Run Stuff Rate: 18% —> 15%
GT Havoc Rate: 18% —> 15%
What in the world happened to this unit? Let’s look to the numbers to help us tell the story of the season and try to figure out where we go from here.
Unproductive Up Front
2021 seemed like the year that we could finally see consistent disruption from the Georgia Tech defensive front. From the arrival of Keion White to the promise of Jared Ivey to the returning production of Jordan Domineck, it was an easy story to tell. Instead, we got more of the same. The entirety of Georgia Tech’s defensive line account for 28 tackles for loss for the whole season; Alabama Edge Will Anderson had 33.5 by himself. In addition to those TFLs, the entire DL rotation put up a measly 3 pass breakups, 11 quarterback hits, and 6 forced fumbles.
Of course, a comparison to the best Edge player in the country isn’t always instructive, but the total landscape tells a bleak story. The GT havoc and run stuff rates checked into the bottom quartile of all P5 teams.
Run stuffs (runs that go for no gain or a loss) and havoc plays (any tackle for loss, forced fumble, pass broken up, or interception) disrupt opposing offenses, either taking the ball outright or setting them behind schedule for moving the chains. The scatter plots below show just how strong of an impact that havoc rate and stuff rate have on winning football games.
The front anchored a defense that was unable to stuff runs or generate havoc, and that meant a defense that didn’t perform. Look at how significant that lack of disruption ends up being for the overall game outcomes. Havoc rate shows an R Squared value of 0.44, while run stuff rate proves even more significant at 0.65.
Let’s look at one more way that we can assess the lack of productivity up front.
Though the correlation is not quite as strong as what we saw with havoc and run stuff rate above, there is a clear effect here as well, as getting more pressure lowers the efficiency of opposing passing attacks. The lack of disruption that GT was able to create allowed opposing offenses to throw the ball effectively.
Defensive End is one of those “or” positions, which means one elite player can vastly change performance. None of the candidates we expected stepped up. White missed most of the season with injury. Ivey got injured just as he was coming on. Domineck didn’t take the steps forward we expected. The defensive tackle position provided a mostly anonymous rotation of mediocre play.
The defensive failings started with a lack of productivity up front.
Boom or Bust at Linebacker
In the middle of the defense, we can glean some insight from tracking stats that David Hale shared in conjunction with evaluating candidates for his All-ACC teams. We didn’t use any of his metrics above because, well, he didn’t include a single GT DE or DT in his considerations, which tells you plenty.
At linebacker, there were 27 players on Hale’s list, two of whom were from GT. Below are some select metrics that give you an idea of what happened with the GT linebacker play.
Quez Jackson
- Ranked 24/27 in pressure rate on pass rush snaps
- Allowed completions on 100% of pass targets for 10.3 yards/target (ranked 21/27)
- Tied for 1st out of 27 in run stuffs
Charlie Thomas
- Ranked 25/27 in tackle %
- Ranked 23/27 in run stuff rate
- Ranked 10/27 in pressure rate (but towards the bottom in # of pass rush snaps)
- Ranked 4/27 in yards/target in pass coverage but had almost 200 fewer coverage snaps than Jackson
Quez Jackson created run stuffs but couldn’t generate pressure or provide coverage on passing plays. Charlie Thomas got lots of pressure and held up in limited coverage, but he missed tackles and couldn’t stuff the run. You might think these complimentary strengths would balance out well on average, but instead it was the complimentary weaknesses that won the day. Here’s how PFF graded out GT’s linebacker play for the year:
Thomas appeared to struggle with injury following the UNC game, but the deployment at the linebacker position still baffles. Eley and Jackson were wildly ineffective, especially in obvious passing situations, and should not have ended up with the lion’s share of the snaps in the middle.
Bust in the Secondary
As we covered above, the inability of the front to create disruption certainly hurt Georgia Tech’s ability to play efficient defense this past season. I wanted to get an idea of how to parse out blame between the front end and the back end. Below, you can see a plot comparing each P5 team’s defensive havoc rate with their defensive EPA/pass. There’s a clear correlation between the two; that checks out, as increased havoc should lead to better results against the pass. But Georgia Tech is farther below the line than any other school in the country.
We can reasonably attribute that difference to coverage in the defensive backfield. Tech’s performance against the pass was far worse than would be expected, even based on the subpar ability to generate havoc. On top of the lack of disruption from the front came poor deployment of players in the secondary and terrible coverage technique.
We can thank David Hale once again for providing us some individual numbers that give color to that poor play. Overall, Hale looked at 35 corner and nickel backs, as well as 20 safeties.
Tre Swilling:
- Ranked 32/35 in Completion Percentage allowed
- Ranked 34/35 in yards per target allowed
- Ranked 32/35 in pass TDs allowed
- Ranked 35/35 in completions of 20+ yards allowed
- Ranked 31/35 in Passes Defensed/target
- Ranked 33/35 in raw QBR allowed
Juanyeh Thomas:
- Ranked15/20 in Completion Percentage allowed
- Ranked 19/20 in yards/target
- Ranked18/20 in pass TDs allowed
- Ranked 17/20 in 20+ yards completions allowed
- Ranked 20/20 in Raw QBR allowed
Once again using Eric Eager’s framework, the defensive backfield is an “And” position. Failure by one player allows touchdown. Over and over again, the three most senior members of the defense were exploited deep. Georgia Tech’s best player in the secondary (Walton) came out to only around average; the rest were catastrophic. It was clear from game one that Swilling, Thomas, and Carpenter couldn’t cover down the field. Clearly, there were no outstanding options available. But the fact that all three of them were still playing the majority of snaps at their positions as the year came to a close was coaching malpractice. Obviously, the secondary coaches have been let go. But couldn’t there have been an edict from the top that we were moving in a different direction for the final quarter or third of the year? There was no upside to continue deploying those guys in the same way, but nothing changed.
As the year progressed, opponents realized just how massively susceptive GT was to throws further down the field.
For Tech’s defense, throwing the ball further down the field was about twice as impactful on efficiency as the national numbers would show. The woeful personnel and technique on deeper throws cost GT again and again.
How bad did things end up?
Worst in the country bad when opponents threw the ball! New Mexico State level bad when we look at EPA/pass and run stuffs together!
Unproductive up front. Boom or bust in the middle. Busts all day on the backend. That completes the story of the 2021 GT Defense. What has to happen for things to look any different in 2022?
Building Block #1
Someone breaks out up front. Lack of disruption continued to be the baseline norm for the GT defense. Havoc rate, run stuff rate, and pressure rate were all at below average levels, and the GT defense paid a price for that, giving up much more efficient offense when those rates went down. A healthy Keion White or a fully available Jared Ivey appear to be the best two bets for finally giving GT a breakthrough in disruption in 2022.
Building Block #2
The Georgia Tech defense has two new coaches, covering the secondary and the linebackers. The head coach has promised to be increasingly involved, and it appears that will include calling the actual defensive plays. The hope here is that the teaching that can happen in the linebacker and secondary rooms will mitigate the worst of the coverage busts we saw this year.
Building Block #3
There are young players with potential in the back 7 of the defense. Deployment choices put members of the secondary in situations where they had no chance of holding up in coverage, and technique issues left massive holes in deep middle zones that were exploited increasingly as the season progressed. The primary players who were deployed incorrectly are all gone, and there’s hope that the younger players responsible for pass coverage will be moldable by the new staff members. My hope is to see Khatavian Franks, Trenilyas Tatum, and Tyson Meiguez get every chance to show themselves next fall, even though the starting MLB returns. Jaylon King was GT’s highest graded safety last year; Walton, Watson, Johnson, and Walker all graded better than the departing cornerbacks. The defensive coaches were afraid to turn the page from the established returners in the 2021 secondary, and they paid the price for it.
New teaching voices, the possibility of a breakout up front, and the loss of the players who were seemingly grandfathered snaps clear the way for something new in 2022. There’s a lot of building to do, but at least some of the obstacles to that building have been removed from the picture.