S&P+ gave Vanderbilt a 56% post-game win probability based on an advantage in yards per play, more scoring opportunities, and perceived poor turnover luck. An efficiency edge and slightly better performance converting scoring opportunities (even though it was ugly) saved Notre Dame is this one.
Confused? Check out this advanced stats glossary.
Vanderbilt’s defense is a still more unknown than not at this point – they beat up on a couple cupcakes early, then were up and down against the Irish. But this felt like a strong bounce-back effort from the Ball State mess for the Irish offense. Notre Dame ran the ball on 65% of its offensive plays, including 19 attempts for Brandon Wimbush. It was a solidly efficient effort both running and passing, including the best performance on passing downs of this young season.
A key to this efficiency, though, was to avoid facing many long passing downs. The Irish ran the ball on 26 of 33 first downs, averaging 6.1 yards per carry and a 50% success rate on those carries. Against Michigan and Ball State the Notre Dame offense faced an average of close to nine yards to gain per third down attempt. The improved rushing attack allowed the Irish to stay “on schedule” against Vandy, with an average 3rd down distance to go at 5.4 yards.
The defense began the game with a lights-out effort, allowing a total of just 36 yards on the first four Commodore possessions, including three consecutive three-and-outs. On those possessions Vanderbilt had just a 17.6% success rate. Afterwards it was tougher sledding – whether by virtue of adjustments, fatigue, or Kyle Shurmur making some pretty strong throws. The run defense was stout, and combined with an early lead led Vandy to run only 25 times all game. Making your opponent fairly one-dimensional is always good, although the pass defense was a little less consistent than ideal.
Hard to factor into the stats again is the flow of the game. For a third consecutive game the Irish built an early lead, and then went flat on offense. Diagnosing the problem is impossible – is it conservatism with the lead and good defense? Issues once scripted plays run out and Wimbush has more to deal with? Early on it felt like the Irish offense had an identity – mixing in runs to the perimeter, with Wimbush as a threat, and taking calculated, low-risk shots in the passing game. That wasn’t the case later (and with the weird Ian Book packages), and hopefully is fixable.
With an uptick in efficiency came a downturn in explosiveness for the Irish offense/ What caused this? My best answer is mostly….nothing. Again, we have good evidence that explosive plays are mostly random, and are best created by just being efficient. If you don’t buy that (but you probably should), you’d also likely point out there are personnel issues involved, and it’s clear right now that the Notre Dame skill position players are more Jose Altuve (singles and doublers hitters) than Giancarlo Stanton. The Irish tried to manufacture explosive passes in the first two games with long downfield passes, with mixed results.
Dexter Williams could fix some of this when he comes back, especially when paired with Wimbush, who is due for some longer runs as well. Still, I would worry far more if the Irish had been less efficient on the day – being this efficient without breaking one or two longer plays has the smell of some bad luck that should regress to the mean in a positive direction.
Notre Dame’s havoc rate ranks 38th in FBS through the first quarter of the year, a nice little jump from finishing 66th last season. The way the havoc is coming is interesting though – the pass rush hasn’t quite made the leap I’d hoped for, although some missed holds and close calls probably undersell the defensive lines effort. But the defensive backs, deflecting passes and finding picks all over the place, are fueling the effort, with a havoc rate as a unit in the top-10.
It was also a huge relief to see the Irish offensive line bounce back from the Ball State performance in terms of allowing disruption. After making the Cardinals look like Bama in terms of havoc rate and run stuffs, Jeff Quinn’s group allowed zero sacks and few tackles for a loss. This isn’t an offense built to play from behind on the scoreboard or behind the chains too far.
Finishing Drives, Field Position, & Turnovers
Converting drives really swung this game in many ways – Vanderbilt created more scoring opportunities, but had two extremely costly turnovers to pair with a missed field goal and the late fourth down failure. The Irish barely scraped out an advantage in points per scoring opportunity, hitting three field goals (although a late Yoon stung) but at least ending these drives with a kick instead of a turnover.
With a better performance in the red zone and converting these opportunities, I think a lot of the fretting over the offense after this game goes away. The Irish averaged around 5 points per scoring opportunity last year, so just a normal 2017 performance would have led to around 30 points and a comfortable win.
The Irish also won the field position battle despite Troy Pride’s interception starting a drive at the Irish 1. Tyler Newsome, as noted everywhere, was exceptional. John Doerer was awesome on kickoffs, and Michael Young showed some great burst on a 48-yard kickoff return. This should be a strength for the Irish, although it’s somehow resulted in mediocre field position at best in recent years.
Once again S&P+ observed some turnover luck for the Irish, who recovered two of three fumbles (including the final lateral play). On film this didn’t feel lucky though – Notre Dame forced tight throws and created strips while protecting the ball very well. The Irish quarterbacks had only three passes deflected, compared to eight breakups of Shurmur. This is a case where the “turnover luck” S&P+ is observing right now is probably a little overstated, but that’s ok.
Reasons for Excitement, Reasons to be Discouraged
#1 An offensive identity seems not too far off?
Kelly and Wimbush have made comments repeatedly about this offense still searching a bit for an identity. But it feels like the approach against Vanderbilt represents what the offense should look like – a run-first attack, with lots of Wimbush runs, sprinkling in play-action to keep the defenses honest (despite 48 rush attempts, the Irish attempted just five play-action throws Saturday). The running attack shown early against Vanderbilt should be the norm – lots of options the defense has to account for, including jet sweeps and fakes that stretch the defense horizontally. If the run attack can be as efficient as it was against the Commodores, especially avoiding long third downs, it should get the offense moving sufficiently against the vast majority of opponents.
As Eric pointed out in his review, the production thus far hasn’t been getting it done, averaging just over 5 yards per play. But I wonder how much of this is the result of conservative play-calling and reliance on the defense with a team that hasn’t trailed yet in this young season. There may be a realization coming that the Irish can’t stop attacking as much (although against Ball State, there was continued attacking that went awry),
Before the year there were a lot of comparisons to 2012, and early in that season there were a lot of games that fit a similar script to what we’ve seen in 2018. Against Purdue the Irish played down to their competition and gained just 5.01 yards per play. Then against Michigan State and Michigan (4.55 and 4.78 YPP respectively), Notre Dame led from the start, and Kelly was content making opponents score enough to climb back versus making a mistake with a young QB.
Now, that stretch of three games was sandwiched by a stomping of Navy (7.10 YPP) and rolling Miami (7.62). We need to see a performance like this, and Wake Forest presents that opportunity, having just allowed Boston College to rack up 7.59 yards per play. As Brian Kelly said this week, “We haven’t attacked the defense at all times and we have to score touchdowns. We have work to do”.
#2 Wimbush has been fine with intermediate to downfield throws
We know the Wimbush weaknesses at this point. The short passing game is not his forte. If the Irish offense wants to move to a game that’s more balanced between running and a short-passing, RPO heavy scheme, Ian Book should be your guy. But if you are truly going to build an offense around Wimbush’s abilities, the intermediate and deep passing game are going to be the most important pieces.
And in those areas Wimbush has been fine, completing deep passes that led to the early lead against Michigan and performing well in the intermediate areas against Ball State and Vanderbilt. He had 11 completions of between 17 and 31 yards against the Cardinal, and followed it up completing 7 of 9 passes for 112 yards on intermediate throws (6-20 yards) last weekend. His third down performance was bad, but Wimbush also was strong in situations that prevented third downs – with successful throws on 5 of 8 situations of 2nd and 8+.
#3 The defense is carrying a heavy load
The Irish defense is ranked #5 currently in S&P+, and has been solid without dominating. Disruption has taken a slight uptick but the stats shows a team closer to the 2017 defense than one taking a leap forward. The potential is there for that leap still to happen – watching the games, this still feels like a better pass rush, better safety play, and better performance at corner. There are series where the defense does dominate.
But there’s also a huge burden on them with an inconsistent offense, and a lack of depth that is concerning. Jerry Tillery and Te’von Coney and Drue Tranquill are doing everything we hoped they would do but also never coming off the field. Already some key pieces from the two-deep like Shaun Crawford, MTA, and Jeremiah Owusu-Koramoah are lost for the year. The top performance level of this defense is in a top-5 to top-10 range, no doubt. I just worry about sustainability over the year, and I think we see some of that even play out game to game as players wear down a bit.
The numbers also see the defense performing well in ways that probably won’t last forever. Holding opponents to 2.69 points per trip inside the 40 (8th nationally) is a huge testament to a great bend but don’t break effort, but also opponents screwing up field goals. The Irish are allowing a sizeable opponent success rate (68th in opponent efficiency) but not really getting burned by big plays (20th in opponent explosiveness), which will change if not patched up.
#4 This Ian Book package is dumb
An initial disclaimer: college coaches have to manage things well beyond my comprehension. 18-22 year olds are difficult to manage and develop, and maybe that plays a role here. I understand wanting to give the current backup QB, who is close in caliber to the starter, a significant role where he knows he will play and contribute every week.
That out of the way, man is the Ian Book “blue zone” package dumb. (If you missed it, BK explained that he defines the blue zone as within five yards of the goal line). Last season the Irish were a top-25 team in terms of converting scoring opportunities, and that was despite a big drop-off at the end of the season. Brandon Wimbush, who had more rushing attempts than any other player in the red zone last year, played a significant role in that. Moving the ball and scoring are harder as the field gets shorter, so why take out a guy who forces the defense to worry so much about the run?
Even throwing out that simple move, there’s good evidence that bringing in these heavy tight-end packages at the goal line and in short yardage is a bad idea. Warren Sharp, an excellent person to read / follow if you care about this kind of stuff, did an excellent job looking at the NFL numbers in these situations. This whole article is awesome and worth the read, but the highlights as it applies here:
- Rushing is more effective than passing in the red zone and goal-line situations
- Runs are more effective out of formations with 2+ wide receivers. Sets with zero wide receivers had the lowest goal line success rate
- Similar NFL data has shown that run success is largely determined by number of defenders in the box, and that when it’s even (blockers vs defenders in the box), more blockers become a disadvantage.
Rushing YPC by Defenders in the Box. Each curve represents Offensive relative rushing advantage. Positive numbers mean more blockers than defenders in the box. negative mean less blockers than defenders.
As the the number of MIB grows, extra blockers become a disadvantage. pic.twitter.com/2uqkb0hIOT
— Josh Hermsmeyer (@friscojosh) June 6, 2018
It’s not apples to apples applying this data to college, but I’m pretty confident the larger takeaways should still apply. Running out of more spread formations should be more effective than these jumbo goal-line packages, and Wimbush is a better runner / gives the defense more to account for.
One quarter down, 3-0
In some ways, this team is off to a start that’s like the anti-2016 team, which was quality but lost a ton of close games. You may remember that team finishing 26th in S&P+ despite the 4-8 record, with 7.2 second-order wins. Those second-order wins take into account what “should” have happened in each game, attempting to take luck and high-variance components of the game like turnovers and look at what’s more sustainable.
At 3-0, the Irish currently have 1.8 second order wins, thanks to a couple victories in close games that arguably could have gone either way. S&P+ gave Vanderbilt a 56% win probability looking at the final stats of this game – one where the YPP edge went to the Commodores, along with more scoring opportunities and some bad turnover luck. These numbers echo what a lot of Notre Dame fans at this point feel – that this team is maybe a bit overrated (18th in S&P+) and due for a letdown any week now.
While by no means is this team a finished product, even if that is the case, I’d argue we should embrace it. These weekly reviews are a lot about the numbers and projecting things moving forward, but speaking from the heart instead of the brain for a second, how fun would it be to scratch out a bunch of close games with a clutch defense, streaky offense, and lethal kicking combination? Would anything make the Notre Dame-hating college football world madder than an Irish team like, say, 2014 FSU, that just keeps winning close games and getting hype despite not being great? I’m all for winning these ugly games, maintaining a spotless (or close record), and having Kirk Herbstreit putting us on upset alert every week.
And without getting too optimistic, the schedule makes something like that somewhat feasible. Stanford is at 27th in S&P+, and somehow is not any more efficient and less reliant on Bryce Love big plays than last year. USC is 39th, 1-2, and still is coached by Clay Helton. Virginia Tech’s opening week win suddenly looks a lot less impressive, and they are ranked 41st. Northwestern received preseason top 25 votes and just lost to Akron. I would love to see Notre Dame come out and dominate against Wake Forest, but if they win by three or something it will be time for full-on Trolltre Dame, the top-10 team everyone thinks is overrated but just keeps winning.