Advanced stats preview: Ohio State’s biggest challenge of the regular season

Penn State week is finally here.

We’ve got a top-ten matchup in primetime, with GameDay visiting, and one of the biggest white outs in recent memory.

The Nittany Lions started off the year on legit upset alert, with a 45-38 win over Appalachian State. Their S&P+ win probability was just 60 percent—if they were to play the Mountaineers again. Since then they’ve played three more-or-less cupcake opponents (Pitt, Kent State, and Illinois), all of whom are ranked 75th or worse in the S&P+. When dealing with the S&P, there will be a standard error, especially if not a lot of data is present. Since the standard error for a team’s quality is so high in Week 1, as they were taking on App State, we really didn’t know how good Penn State was going to be this year.

But, the opponent-adjusted advanced stats still like Penn State: they’re sixth overall and fifth and 22nd in offensive and defensive S&P+, respectively. This should be Ohio State’s most difficult game of the regular season. Their overall profile is of a team with an excellent offense— especially on the ground— that leans towards efficiency over explosiveness (think 2016 Penn State). They might also have some questions on defense, and they have a senior quarterback who has been far less consistent than he was last season.

---

Top level stats

If you’re unsure about the definitions for any stats, check out the advanced stats glossary.

---

OSU Offense vs. PSU Defense

These charts are intended to help visualize relative strengths and weaknesses. The farther apart the two teams’ radar points are from each other for any given statistic, the more lopsided that matchup is expected to be. The closer to the outer edge of the radar, the better. Here’s the above data in table form:

---

PSU offense vs. OSU defense

And definitely check out the full team advanced stats profiles for both teams.

Alright, let’s get to the questions that will decide the game:

---

Can the Buckeyes get out to an early lead?

So far, Penn State has been a second-half offensive team:

• vs. App State: tied 10-10 at the half. Needed overtime when App State scored 4 touchdowns in the fourth quarter to tie.
• vs. Pitt: led 14-6 at the half, before scoring 37 points in the second half.
• vs. Illinois: trailed 24-21 in the third quarter, before scoring 42 points in the second half, including 35 in the fourth quarter.

These facts say two things. First, a sluggish start for the Penn State offense could allow Ohio State to mount an early lead. Second, Penn State’s offense is obviously capable of absolutely exploding in the second half, so no potential lead should be considered safe.

The counterpoint is that Penn State hasn’t come anywhere close to playing a team that is its equal or better talent-wise, and talent differentials probably account for a lot of late-game offensive explosions.

Without diving too deep, it looks like McSorely might be a big reason for the Nittany Lions volatility by quarter. In the first three quarters, he’s completed between 50 to 56.5 percent of his passes for 162 to 208 yards. In the fourth quarter, his completion percentage on the year jumps to 62.1 percent and he has 370 passing yards and six touchdowns, more than the four total in the other three quarters.

---

Can Ohio State establish the run game vs. Penn State?

There’s been somewhat of an existential crisis in the fan community about the state of the OSU run game.

Against TCU, the Horned Frogs actually averaged more yards per rush than the Buckeyes did, with 5.6 yards per carry compared to 4.3. On the season, Ohio State is averaging just 5.4 yards per carry compared to 5.8 last season. Dobbins has only had one 100-yard game, and against Tulane he managed just 55 yards at five yards per carry. Mike Weber hasn’t averaged over 3.9 yards per carry since the Oregon State game.

But the advanced stats tell a different story than the total or basic average rushing numbers. Their opportunity rate, which is the percentage of carries that gain at least five yards, is 57.4 percent, which is way above last year’s rate of 46.1 percent (which was third-best in the country). And their rate of 10+ yard rushes is also better than last year’s, at 16.7 percent of runs compared to 15.9 last season. Obviously we’re only 4 games into the season, but that’s at least some indication that reports of the Buckeyes run game’s death were premature.

Instead of the run game being worse than we’re used to, it’s probably just more accurate to say that it’s just being de-emphasized in favor of the passing game — although not necessarily on purpose.

Here’s Meyer:

The accuracy of the passing and the skill set of our wide-outs,” he said. “We have a big offensive line. We never had a chance to really get the run game going the way we wanted to. You notice every time we got through that safety was right there to get us down, but that created single coverage I was talking about. That’s just -- offensively the accuracy of the quarterback and the kids making catches.”

Essentially what Meyer is saying is that opposing defenses have still been prioritizing defending the run, creating single coverage that Haskins and the wide receivers have been able to exploit.

The numbers back that up. The best stat to capture that is probably standard downs run rate. (Standard downs are first downs, second-and-7 or fewer, third-and-4 or fewer, and fourth-and-4 or fewer — essentially the downs where the offense has an option to run or pass.) Ohio State ran on 60.9 percent of standard downs last season, which was the 52nd-most in the country. This year the Buckeyes are running on just 56.6 percent of standard downs, which is 76th in the country — and that’s with garbage time removed.

When opposing teams stack the box and outnumber your blockers, you can re-equate the numbers by either using the quarterback as a ball carrier and reading defenders, or you can pass more. Making the quarterback run has an extremely high floor for your offense, but there are a few downsides.

First, depending on the quarterback’s skill set, the defense might be able to dictate the quarterback’s reads, forcing the quarterback to keep the ball if the running back is a bigger threat. This happened to J.T. Barrett last year against better defenses — the opposing defense would force Barrett to run instead of handing off to Dobbins.

Second, even though running the ball has a high floor, it also generally has a lower ceiling. From Bill Connelly’s analysis of 2017 running backs:

As with what people have begun to firmly establish on the pro side — I highly recommend picking up a copy of the Football Outsiders Almanac 2018 and checking out the “establishing the run” essay by Ben Baldwin — running is a reasonably lower-ceilinged endeavor. It’s lower-risk, too, and some teams have certainly figured out how to run more than others, but for a majority of feature backs, handing them the ball was likely to put you behind schedule.

It was also far less likely to produce big plays — only 18 of these 83 [running backs with 150+ carries] produced a marginal explosiveness above plus-0.0 points per successful run.

That suggests that the 2018 Ohio State offense could have a higher ceiling than the Buckeyes’ 2015-2017 offenses because of their increased ability to pass.

---

What happens if the Penn State defense prioritizes defending the pass?

So what happens if Penn State goes with a different defensive strategy against Ohio State and decides to prioritize defending against the passing game, even if it opens some holes for Dobbins and Weber?

Honestly, we don’t really know. TCU brought pressure a good bit and wasn’t really all that successful at getting to Haskins. But, Penn State could also decide to combine blitzes from varying defenders with dropping eight, trying to confuse Haskins’ reads. We don’t really know how Haskins would respond. However, he has just one interception this year, and this is the kind of close matchup where any turnover could decide the game’s outcome.

But, it could also open things up for the run game. As I said above, the run game has been solid this season even if defensive looks have led to more passing. So there’s a world out there where Ohio State’s run game goes off against Penn State because the Lions decide to prioritize pass defense.

---

The Buckeyes have been solid at protecting Haskins. Will that continue vs. Penn State?

Penn State’s pass rush is solid, especially when it really needs to be. For example, their overall sack rate is 8.1 percent (34th), which is good, but not elite. But on passing downs that percentage jumps to 13.7 percent (15th), and on open play blitz downs, their sack rate is 23.1 percent (3rd). Blitz downs are defined as first-and-18 or more, second-and-14 or more, and third-and-3 or more.

TCU was by far the best, and most aggressive, defense that Ohio State has seen this season, and they managed just one sack during the game. The offensive line has allowed just a 2.7 percent overall sack rate (23rd), and they’re 16th in passing downs sack rate.

While Ohio State’s blitz down sack rate ranking is low — 89th — Ohio State has still only allowed six total sacks this season, and most of them have been against Tate Martell.

Maybe the most interesting stat is the blitz down success rate, where Penn State drops to 60th, despite being third in actual sack rate. That suggests that Penn State hasn’t actually been all that successful on blitz downs even though they’ve gotten a relatively high percentage of sacks — it’s not just important whether you get the sack, it’s also important whether the blitz disrupts the offense’s play, and the Nittany Lions are just average in that area. Conversely, Ohio State is actually excellent against the blitz, with a 46.2 percent blitz down success rate, which is eighth-best in the country.

That said, Ohio State’s offensive line definitely still has to worry about Shareef Miller (three sacks and six tackles for loss) and company because a) we’re working with limited data still, and b) this is probably the best defensive line that the Buckeyes will have seen.

---

Can the defense stop Miles Sanders?

While I’m very interested to see how the Penn State defense decides to defend against Haskins, the question of the game may be how well Ohio State can stop Penn State’s new starting running back, Miles Sanders.

It’s tough to follow a first round NFL draft pick in Saquon Barkley, but Miles has so far risen to the occasion. It’s crazy to think, but Sanders was actually a much bigger recruit than Barkley. He was the top-ranked running back in his class, actually. The Penn State rushing offense is 12th in rushing marginal efficiency and fifth in opportunity rate (which is better than Ohio State’s).

Sanders’ personal opportunity rate is 64.8 percent, his marginal efficiency is 15.5 percent, and his marginal explosiveness -.01. For comparison, Barkley was 39.2 percent, 1.1 percent, and .08. We have to keep in mind that Sanders hasn’t seen nearly the same kinds of defenses that Barkley saw throughout the entire season last year, but those numbers still describe a high-efficiency back that has lower explosiveness than Barkley — which is what most observers tend to see, too.

Ohio State’s issue this season has consistently been the big play. They’re only allowing five+ yard runs on 35.7 percent of opponent runs (12th) and stopping runs for no gain or a loss on 27.4 percent of runs too (13th). But they’re also 53rd in rushing marginal explosiveness, notably allowing the 93-yarder against TCU, but also a few against Oregon State, too.

Sanders is certainly capable of big runs, but the bigger threat with him is probably just him falling forward for five yards at a time, until the eight-10 yard runs come more frequently later in the game. So, I’ll be watching whether Ohio State’s linebackers set the edge and fill gaps early on, and whether the defensive line can rotate enough so that they’re not tired and getting blown off the ball in the fourth quarter, because Penn State’s offensive line has certainly improved.

---

Will Ohio State’s pass rush still get to McSorely?

This is partly a life-without-Bosa question, and partly just an understanding that the Nittany Lions offensive line is actually solid.

They’re giving up sacks on 3.1 percent of passes, which is 30th in the country. They’re worse on passing downs, allowing sacks at a 6.7 percent rate (61st). That should be a favorable matchup for Ohio State based on the numbers — the Buckeyes are ninth and third in those categories.

Besides applying pressure, it’s worth noting that McSorely’s numbers are down by a good margin so far this season. Even though App State is the best defense they’ve faced, McSorely is completing 53.8 percent of his passes and is averaging just 6.8 yards per attempt with a marginal efficiency of 4.7 percent (Haskins’ numbers are 75.5 percent completion rate, 10 YPA, and 20.7 percent marginal efficiency).

The defense will certainly still have to be concerned about his passing over the middle of the field, particularly to sophomore KJ Hamler, who has really come on as an explosive option, though.

---

If Penn State creates a scoring opportunity, can Ohio State keep them from getting a touchdown?

The final big defensive concern is related to Ohio State’s ability to prevent Penn State from finishing drives. Penn State averages 5.61 points per scoring opportunity (20th), while the OSU defense allows an average of 4.06 points per opportunity (46th).

The Nittany Lions have scored on 100 percent of their red zone opportunities, and touchdowns on an astounding 95.5 percent of them — all but one of their 22 red zone trips. Ohio State hasn’t allowed many red zone trips to begin with (just six), but that’s something to keep in mind.