Advanced Stats Preview: Clemson vs GT

Robert’s asked me to take the column for a spin this week. Will it be just as good as his analysis? Here’s hoping!

Let me be clear: Climpsun/Clemzin/Clemzun/Clemson is a juggernaut. There’s no two ways about it. They are the obvious class of the ACC and, in most recent years, have faced no formidable threat from the rest of the conference (this year’s future Notre Dame shenanigans aside).

In keeping with that, most, if not all, metrics have Tech losing this game, oftentimes by a bunch. Notably,

• Vegas: Clemson -27; O/U 59.5
• SP+: Clemson -22
• FPI: Clemson win prob 95%

I don’t say this to be depressing or anti-Tech — I just want to underscore the context and set expectations here.

But we lean on the advanced numbers to find glimmers of hope — in the face of overwhelming football domination, where can we find success? Let’s find out together.

The Data Dump

I don’t have Robert’s fancy EPA numbers hooked up into my data pipeline yet, but here’s what I do have for you:

What 2 Watch 4

When Clemson has the ball

It doesn’t matter what Clemson does with the football; they’re typically good at everything. What really sets them apart from other teams is their ruthless creation of, and efficiency during, scoring opportunities (defined as drives that pass the opponent’s 40-yard-line): the Tigers are getting in this pre-red zone (we’re really going to have to come up with a name for this) on 62% of their drives and scoring 81% of the time when they’re in said zone. Considering these stats together, Clemson is scoring on ~50% of their offensive drives. This is an insane clip, and it gets even more absurd when you consider their points per scoring opportunity — they’re not just plain scoring every other drive, but more often than not, they’re coming away with a touchdown.

What can Tech’s defense do to stop the unstoppable machine? It’s a bit difficult to say simply based on the numbers. It’s entirely possible that Tech’s best chances to stave off being ground into dust are 1) to force turnovers (IE: increase its havoc rate), 2) contain Travis Etienne (IE: maintain its rush success rate allowed), and 3) in accordance with #2, get push against a young Clemson offensive line (IE: maintain its overall success rate allowed and line yards allowed — the latter of which is not shown above). As with everything in this game, these three keys are obviously easier said than done, but we saw the Georgia Tech defense do two of these three things last week: Tech got very aggressive with ripping the ball out of Louisville RBs’ hands, and the Tech defensive line bullied the Louisville offensive line in spots last week. This week, talent differential aside, there is a chance to work Clemson’s OL — just like Miami did (despite their overall disappointing performance) — and at least make Trevor Lawrence think about where and when he’s throwing the football. If Tech can keep the pressure up on said line, it’s possible they generate more “havoc” plays (INTs, PBUs, forced fumbles, tackles for loss). Additionally, if the Tech defense is able to generate said pressure, then it’s likely that they can put together a good rush defense performance, like they did versus Louisville. Lots of hypotheticals, of course — but all that matters is that it could happen.

When Tech has the ball

If there are holes in this Clemson defense, they’re against the run — the Tigers allow 0.110 EPA/rush (#63 in the nation out of 76 active teams) — and on third down, where the Tigers allow 0.453 EPA/play (#64 in the nation). Notably, third downs only make up 25.5% of the plays they’ve faced (#2 in the nation), signaling that they frequently hold opponents to three-and-outs (if the success rate numbers didn’t give that away already). If Tech’s offense is to have any success moving the ball against the Tigers (analytics nerds, cover your ears now), it’s going to have to rely on its stable of running backs to carry the day and, in a more difficult vein, it’s going to have to sustain drives versus a stout defensive front. Once again, these are tasks easier said than done. Common sense dictates that Clemson also recognizes this mismatch, so expect them to key on the run and force Tech to throw the ball, much like their (and many other teams’) gameplan looked last year. However, Tech has a more effective passer under center in 2020, so it’s possible that Jeff Sims could find and hit seams in the middle of the field if/when Clemson stacks the box against the run (although, granted, this plays right into Clemson’s defensive stance — 29% pass success rate and -0.248 EPA/dropback allowed).

Prediction

Before the season, our in-house model picked Clemson by ~22 points and gave them a ~90% win probability (I specify before the season because I haven’t had a chance to tweak things more for 2020). That’s right in-line with current SP+ and just a bit off from Vegas (if you’re into their persuasions).

It’s hard to disagree with those numbers. Like I led with, Clemson is the class of the conference and a Playoff favorite; it was always going to be difficult to get a jump on them. But if Tech can get its ground game going when they have the ball and generate some chaos when they don’t: