2018 Notre Dame Advanced Stats Preview

It’s not surprising that advanced stats see the Irish as a top 10 team entering 2018. After a significant leap forward on defense last season, Notre Dame returns more production on that side of the ball than almost every team in the nation. The offense, which fueled by a dominant run game performed like a top-10 unit for much of 2017 before a November slowdown, loses far more but as usual has blue chip options waiting in the wings. So how will Notre Dame’s explosiveness, efficiency, look this season? Below in the 2018 Notre Dame advanced stats preview you’ll find ten key questions that will determine if Brian Kelly’s team will be a playoff contender or set up for an underachieving year.

*Semi-required reading here is Bill Connelly’s always-excellent preview, which was published in May. This preview will heavily utilize Bill’s S&P+ numbers, including this advanced stat profile. I’ve tried to provide the background to define the advanced metrics used, but if you run across something unfamiliar this advanced stats glossary is a helpful resource..

The Offense

Revisiting 2017 

Let’s quickly take a look back at Notre Dame’s offensive performance in 2017. Fueled by an explosive running game, the Irish offense spent most of the year in the top 10 of Offensive S&P+, only to fade in November to finish 26^th nationally. The run game accounted for almost all of the heavy lifting, finishing 5^th in Rushing S&P+ versus 64^th in passing. The Irish run game was well rounded, finishing 18^th in Rushing Success Rate (Efficiency) and 5^th in rushing explosiveness (as measured by Rushing IsoPPP).

It was extremely effective except when it wasn’t – over the last four games of the season, beginning with Miami, a solution to the Irish run game began to form. Defenses stopped honoring the threat of the pass or jet sweeps and began keying on Brandon Wimbush and Josh Adams, daring the Irish to beat them another way. Wimbush, who struggled all season with consistency, footwork, and accuracy, appeared to have his mental challenges as well turning things around when struggling to complete basic passes.

These splits come with an obvious caution – the last four games of 2017 include Miami (the good one) and LSU (although they were missing several key players), but take a look at this disparity:

• First nine games: 72% average offensive percentile performance, 41.3 points per game, 6.9 rushing yards per carry
• Last four games: 29% average offensive percentile performance, 18.3 points per game, 4.1 rushing yards per carry

#1 How far does the run game fall off?

It’s a safe bet that the Irish won’t replicate a top-5 rushing performance in 2018. If anything, advanced stats probably underrate the impact of Notre Dame’s losses – Quenton Nelson and Mike McGlinchey show up as losing an enormous amount of experience on the offensive line, not a garage door with arms and his extremely mobile safe leaving as top-10 NFL draft picks.

Maybe it was the disappointing end to the season and injuries derailing his Heisman campaign, but I also think Josh Adams’ absence is still a little underrated. His 10.1 highlight yards per carry, a measure of how well a runner takes advantage of open space, trailed only Bryce Love among P5 runners and significantly outpaced Saquon Barkley (8.3) and Lamar Jackson (7.7). “Well, he had the best offensive line in the nation!” you might argue, and I’d respond by letting you know that according to Pro Football Focus Adams averaged the highest mark nationally in yards per carry after contact. This is not a guy who just got lucky running through open holes.

Still, an extremely explosive attack still feels viable despite these massive losses. One important element that certainly aided Adams was the legs of Wimbush, who was himself Lamar Jackson-like as a runner, albeit with a lighter workload (Wimbush 8.0 yards per carry, 7.3 highlight yards per opportunity, Jackson 8.7 and 7.7 respectively). Dexter Williams figures to miss something like a third of the year, but essentially replicated Adams explosiveness numbers in his limited carries in 2017, with 10.0 highlight yards per carry and 9.2 yards per carry overall.

It will take this third year as offensive coordinator to verify as a trend, but Chip Long may also be a playcaller who can engineer big runs. In his lone year calling shots at Memphis, the Tigers finished 28^th in Rushing IsoPPP and broke 8 runs of 40+ yards (T24 in FBS) despite a pass-heavy attack.

Much will depend on the offensive line, which figures to have a high floor but may experience some growing pains. Hiestand, McGlinchey, and Nelson cannot be replaced. But it’s also not necessary to be that dominant on the line to still be effective, and on paper this combination sounds pretty good:

• Two senior captains with a combined 62 starts that were starters on the Joe Moore Award winning line last year
• Two more starters that shared a starting rotation at tackle last season and were top-100 recruits
• Another top-100 junior who beat out the guys above to win the left tackle job
• If any of those don’t work out, a couple massive top-200 tackles entering their second year, plus a senior interior lineman

I’m not predicting dominance, but this can be a very good line, with a floor that’s probably above average but improves throughout the year. Better decision-making by Wimbush, both in the run and passing game, can go a long way in helping the line as well. A top-40 efficiency mark, coupled with top-20 explosiveness, is definitely in play, and would likely result in another top 25 rushing offense.

I’ve gotten this far without really mentioning the running back division of labor, in part because it’s entirely unpredictable and also because I think it’s third in order of importance behind Wimbush’s running ability and the offensive line. The options run from Williams, whose strengths and weaknesses have been extremely well documented, to Tony Jones (simultaneously the least explosive option and very possibly the best one for ND overall offensive success), Jafar Armstrong, and freshmen Jahmir Smith / C’bo Flemister.

You can throw Avery Davis in there as well, but ultimately I feel good about 1) an explosive option emerging and 2) running back not making or breaking this thing. If I was a betting man I’d be purchasing some Tony Jones stock as undervalued (weird post-’17 hype backlash after a year where he was beat up) and Armstrong as the other player to emerge.

#2 How far can a quarterback, presumably Brandon Wimbush, go in taking the passing game from a liability to something better?

The 110^th ranked passing success rate in 2017 feels about right, doesn’t it? Brandon Wimbush’s passing accuracy will swing this season one way or the other, and if it’s closer to a 2017 remix than turning over a new page, it won’t be his job for long. A 40% passing success rate is average, and Wimbush’s regular season starts the passing success was awful to bad five times (<25% against UGA and Miami, <35% vs Stanford, USC, and BC), around average five times (Miami (OH), Temple, NC State, Wake Forest, Navy), and good (>45%) once, at Michigan State (50%).

There was a lot tied in up those poor performances that’s been covered ad nauseum this off-season. I don’t think anyone has quite untied the interrelated issues, from footwork to reads to accuracy and confidence. But it was evident last year that once things seemed to go wrong in one of those areas, others were impacted and it snowballed quickly, often when the lights were brightest.

Even in Wimbush’s inconsistencies, however ugly things looked at times, were some bright spots. The completed passes went for distance – 13.8 yards per completion would be a top-25 mark nationally, and ND finished 18^th in passing explosiveness (IsoPPP). Wimbush’s running ability, as both a threat, in designed runs, and as a scrambler – helped allow the Irish to run far more than average on passing downs. On these downs (2^nd and 8+, 3^rd/4^th and 5+), the Irish ran 44% of the time, the 16^th highest rate in FBS. Some of this is Wimbush scrambling, but a lot was also confidence running the ball on 2^nd and 10 with a strong offensive line.

This worked well– Notre Dame finished 14^th in Passing Downs S&P+, including 3^rd in explosiveness on those downs. Is that sustainable with such poor efficiency passing, and a weakened rushing offense heading into this fall? Absolutely not. And a deeper dive shows just how much that success was driven by the run versus the pass, potentially when defenses weren’t expecting it (explosive plays here are runs of 12+ yards and passes of 16+):

Notre Dame broke five rushes of 60 plus yards on passing downs, which was probably a little fluky and widens the yards per play disparity. Still, remove those and you still have better efficiency and yards per play (without those runs, rushes average 6.26 YPP) on the ground in passing downs. The passing game will have to be better in these situations, because the run game almost certainly won’t be as insane. There was definitely some randomness going the right direction to last year’s passing explosiveness – it’s really hard to be that inefficient but break that many big pass plays – that is apt for regression.

Wimbush will probably not be as well protected either. The Joe Moore Award winning line was definitely better run-blocking than pass blocking, and that combined with some poor Wimbush decisions resulted in an adjusted sack rate of 65^th in the nation. Replicating that mark might be doable, with improvement in Wimbush’s pocket presence offsetting the downtick in pass protection. Again, there’s a wide swath of possible outcomes, but mobile quarterbacks tend to encourage a slightly higher sack rate as they attempt to escape pressure rather than throw the ball away like their more statue-esque QB coach counterparts.

#3 Where else can the offense improve in Year 2 of Chip Long?

In general, I think the operating assumption is that the offense will be worse than 2017 in most facets, the most likely exception being the pass game. But there’s several areas beyond quarterback where Notre Dame under-performed last year and provide easy room for potential improvement.

First, there’s hidden yards to be gained to make things easier for the offense via special teams. In 2017 the offense inherited the 108^th best starting field position, on average starting at their own 28. By comparison, Stanford’s average drive started nearly 5.5 yards further downfield. Do the back of the napkin math – around 5 yards and 12ish possessions per game – that’s close 60 yards per game lost! The Irish were poor last year in both kickoff and punt returns as well as coverage, and have the legs and athletes to be far better.

Receiver performance, both at wide out and tight end, was inextricably linked to issues at quarterback last season. Bill Connelly’s analysis on returning receiver production shows that losing a lot of production typically bodes poorly for an offense the following year. But taking a closer look at the departing players, there’s clearly room for better performance:

In parentheses above is the number of passing targets for each player in 2017. You can see than Equanimeous St. Brown, by far the most targeted receiver, was among Notre Dame’s least efficient player in a close race with Alize Mack. While targeted far less, Chase Claypool and Miles Boykin provided far better success rates without much drop-off compared to Kevin Stepherson and St. Brown in explosiveness (looking at yard per target or yards per catch). Stepherson was strong, but unreliable with his shadow suspension limiting his contributions.

With a trio of talented young freshmen also pushing for playing time, as well as a wealth of guys who can step in at tight end if Mack doesn’t improve, receiver production could be much better in 2018. There’s a real chance Kevin Austin can step in and be a difference-maker right away, same with Cole Kmet as a sophomore. It’s also worth noting that pass efficiency to running backs was extremely poor last season (Josh Adams gained 4.8 yards per target, Tony Jones 1.5) and could be a much bigger part of what Chip Long wants to do if a quarterback can complete short touch passes.

The last, very non-statistical reason to expect some positive traction is a second year in Chip Long’s offense providing continuity. There was a lot of good in Long’s scheme in 2017, but by all accounts he was limited doing some things he’d like to (more two back sets, screens, RPOs) by lack of reliability at QB. Already we saw an impact on tempo, as the Irish offense went from one of the slower teams in the country (108^th in adjusted tempo) to among the fastest (17^th). Notre Dame may be able to deploy the hurry-up offense strategically based on situation and opponent, but should be even more comfortable actually playing fast in 2018.

And despite the lack of proven playmakers at skill positions, Long’s offense in 2017 succeeded in setting his players up to win in space. Stylistically 68.7% of opponent tackles were solo, ranking among the lowest numbers in FBS (73.8% was the national average, Notre Dame was 25^th lowest in opponent solo tackles).

Defense

#4 How do you make the leap from good to great?

Last season the Irish fielded a good defense – top 25 in Defensive S&P+, with no real holes. Mike Elko’s unit was extremely balanced – top-25 in passing defense, rushing defense, opponent explosiveness, and opponent efficiency. Adjusting those efficiency and explosiveness numbers to account for strength of schedule and the Irish were top ten in both measures in 2017. Building on that success, Clark Lea inherits a unit that ranked 1^st in Bill Connelly’s rankings of returning production on defense. S&P+ projects the defense to be 6^th nationally as a result – so how will the Irish make that leap?

Returning continuity is enormous, and gives Notre Dame a chance to continue strong performance in key areas. In addition to finishing 5^th in Defensive IsoPPP+ (opponent adjusted explosiveness allowed), the Irish allowed just four plays over 40 yards all season (tied for 4^th in FBS) and one over 50 yards (2^nd). The other power five teams in that range all fielded extremely strong defenses – Washington, Alabama, Michigan State, Miami, Georgia, and Ohio State all allowed three or fewer gains of 50+.

A bit of regression is probably likely – as Bill Connelly’s extensive research has shown, explosive plays are mostly a random byproduct of successful plays. The best way to prevent big plays is just to limit any successful plays, which the Irish did very well, but if you look at 40 or 50+ yards yielded year to year you find little consistency, even among the Alabamas of the world. Notre Dame will likely let up a few more big plays, but can replicate much of last year’s success by continuing to limit defensive efficiency.

The room to grow is in disruption – Notre Dame ranked just 66^th in havoc rate, which measures the percentage of plays that end in a tackle for a loss, forced fumble, or pass deflection (including interceptions). This was spread fairly evenly against run and pass defense, with run stuff rates (tackles for no gain or a loss) and adjusted sack rates both ranking very average (stuff rate 51^st, adjusted sack rate 64^th).

For the defense to grow into a dominant, top-5 unit, there has to be more disruption. The Irish can’t count on the success limiting big plays to be quite as stellar this season, and it would be a gamble to assume they will replicate their success limiting opponents scoring opportunities (18^th in opponent points per trip inside the 40). Success with takeaways early in the season (another statistic that’s fairly random year to year) masked some games where the Irish weren’t great limiting opponent success rates, and then resulted in some poorer performances when the turnover well ran dry late in the year. So how can Notre Dame at least maintain, if not increase its ability to limit opponent efficiency? More disruption that will put opponents behind the chains early, and create passing downs where the Irish were already excellent last season despite a middling pass rush.

#5 Can the run game stay stout with all the shuffling?

Last season the Irish finished 13^th in Rushing Defense S&P+ – as mentioned earlier, not stuffing a ton of carries but also very rarely allowing big gains. Against opposing offenses that thrived on the ground– UGA, Navy, Stanford, and LSU – the Notre Dame defense limiting each to under 4.3 yards per carry.

It’s easy to drink the Kool Aid that a repeat performance is possible, with three of four starting defensive linemen returning (and the old starter having been passed up in spring practice) and two of three starting linebackers. The tricky piece of that assumption is that those pieces will perform at the same level from different positions – between the defensive line and linebackers four of the front seven have slid into new roles.

The few losses that ND will sustain also figure to leave something to replace in the run game. Jay Hayes and Andrew Trumbetti each had at least six run stuffs and stronger in run defense. Nyles Morgan had a bit of a disappointing senior season with sky- high expectations, but still led the team in run stuffs with 17. His absence in the linebacking corps could result in a ripple effect – Tranquill moving in from rover and Coney inside leaves big shoes to fill for a new starter.

Still, there is depth to account for those absences. Myron Tagavaialoa-Amosa and Kurt Hinish saw meaningful snaps as true freshmen, and should spell Jerry Tillery and Jonathan Bonner. Tillery could be even more disruptive moving from noseguard to 3-technique, and Khaled Kareem was more productive than Jay Hayes on a per-snap basis. Freshmen like Jayson Ademilola and Ja’mion Franklin could force their way into time beyond the four free trial games they now get before burning their red shirt opportunity.

In theory, the rotating and replacing should all even out at worst. But it will be interesting to keep an eye on, especially against a Michigan offensive line that struggled a bit last year and with Bryce Love also making an early season trip to South Bend.

#6 Could an improved pass rush cement elite status for the Irish pass defense?

In Mike Elko’s first year the Irish rocketed from 86^th in Pass Defense S&P+ in 2016 to 7^th. There was a beautiful confluence of the introduction of a strong, coherent scheme synching with young players, who had taken their lumps the year before, making significant progress. The return of a healthy Shaun Crawford and Nick Watkins to provide experience and depth at corner made an impact as well. Elko had a long track record of excellence on passing downs, and Notre Dame was 4^th best nationally in Passing Downs S&P+ last season, stifling opponents to a 24.2% success rate (8^th in FBS) in those situations.

The enormous opportunity is that the Notre Dame pass defense performed at this high level without much in the way of a pass rush. Elko’s defense succeeded in providing increased pressure without giving up big plays, but still finished just 64^th in adjusted sack rate, including finishing a dismal 98^th in passing down sack rate. The chart below shows the disruption (sacks, TFL, pass deflections, and forced fumbles) of each member of Notre Dame’s defensive line on a per snap basis, with the number of snaps in parentheses.

The pass rush should be better next year, as around 900 DL snaps will transition from two of Notre Dame’s less disruptive ends (especially rushing the passer) to the two that wreaked the most havoc. Trumbetti was often lined up in third and long passing situations, and replacing him with a package that features something like Okwara, Kareem, Tillery, and Daelin Hayes figures to be an upgrade. I was surprised to see Daelin Hayes so low on the list of disruption, so it will be interesting to see if Okwara begins stealing more and more snaps as the season progresses or if a dark-horse candidate like Ade Ogundeji emerges as one of the better pass rushers.

There’s a lot of reasons to believe in an even better performance within the top 10 for the Irish passing defense with Lea this year. Love returns after finishing as a second team All American as a true sophomore, Troy Pride passed Nick Watkins and provides a ridiculous ceiling with his freakish athleticism, and it seems safe that play at safety will only improve, the only question is how much (more on this to come). But an improved pass rush will only help the secondary, increase efficiency, and make turnovers more likely.

#7 Can someone step up at rover to make the linebackers one of the nation’s best?

The floor for the ND linebackers as a group, provided good health, is probably the highest of any position group. As seen above, Coney and Tranquill were the Irish’s two most disruptive players at the position last year, with Coney leading the team by far with a tackle (solo or assisted) on 21% of snaps he played. Tranquill wasn’t quite as productive racking up tackles, likely in large part by virtue of different responsibilities at rover, but was the second leading backer in terms of havoc (tackles for a loss, pass breakup, or forced fumbles) per snap.

The one piece not captured above that gives pause is how the unit will perform in pass coverage. Coney had zero pass break-ups last year, and that’s the obvious area for his development. Tranquill had four, including a pick, and should be an upgrade vs the Coney/Morgan combination.

Rover is the one question mark, with a portfolio of high upside options. Between Asmar Bilal, Shayne Simon, and Jeremiah Owusu-Koramoah, there’s a mix of athleticism, raw talent, and experience that seems to have a high floor (a recurring theme, defensively). Bilal as the favorite should provide strong run support, and if a young guy beats him out that’s probably a good sign for the health of the position. In addition, the Irish may spend a lot of time in nickel, where Shaun Crawford and Nick Coleman figure to be options and may be even more effective than the base defense.

#8 Will the safeties make more plays while still limiting deep passes?

It’s funny how quickly expectations can change. Coming out of the Brian Van Gorder era, the defense desperately needed steadiness and competence at safety. A hallmark of the BVG tenure was watching a safety frantically scrambling to get into the right coverage, recover after biting on a trick play, or taking a poor angle on a tackle that led to a big gain.

In 2017, Irish fans received competence and steadiness, and were still largely dissatisfied. Look, I’m not here to sell you that five total pass break ups, one TFL, and no interceptions is performance worthy of a standing ovation. But the flip-side of the lack of disruption was the lack of explosive plays given up – Notre Dame allowed just one pass play of 40 yards or more last season, best in the nation (and zero allowed of over 50 yards, which obviously also ties for best in FBS). It’s also not surprising that Nick Coleman, a converted corner in his first year the position, and Jalen Elliott, a sophomore with limited experience turned starter, weren’t exactly All-Americans out of the gate.

The bar will be higher this year, and yet there seems to be nothing but potential to build on 2017’s performance. Coleman and Elliott figure to improve and provide a very solid floor for the position, with the possibility of either breaking out (Elliott was a spring game darling, Coleman the unofficial king of Culver). If they get beat out by Houston Griffith or Alohi Gillman for starting roles, that’s probably a good sign for depth and overall elevation of performance at the position. There appears to be four good options, a high floor, and the rest of the defense should be good enough that safety doesn’t have to be spectacular again in 2018 – just a little more disruptive.

Big Picture

#9 What do advanced stats predict for Notre Dame, and what don’t they know?

The hype for Notre Dame as a top 10 team is not undeserved, especially with what should be one of the best defenses in the country. S&P+ projects the Irish #7 nationally, with the 6^th ranked defense and 22^nd ranked offense. The Power Rank agrees and also has ND at 7^th. Preseason FEI ratings aren’t out yet but I’d expect them to be equally optimistic.

While each advanced stats system uses their own methodology, it’s generally a combination of past performance (with variability in how heavily each ranks the recent past versus the last four to five years), recruiting, and returning production. So what questions or concerns are these numbers unaware of?

1. How well does Clark Lea make the leap?
   While there will be some small differences in philosophy, continuity in the Elko 4-2-5 is a much benefit to the ’18 defense. Lea has said and done all the right things so far, but no one knows how his ability to game-plan, call plays, and make adjustments will compare to his mentor. Some rookie defensive coordinators have made the promotion look easy (like Jim Leonhard at Wisconsin), but it wouldn’t be surprising for Lea to take some lumps just as a result of inexperience.
2. Is Brandon Wimbush’s returning production a benefit to the passing game?
   Projections will see the quarterback’s returning production, both passing and rushing, and chalk this up as a benefit for the Irish. And it definitely could be! But if the worst case scenario happens, and Wimbush plays poorly or Book takes over, S&P+ and others can’t see that coming.
3. The schedule isn’t top-heavy, but offers lots of quality opponents with few breaks
   The projections don’t see any one opponent as a top ten team, but several knocking on the door that certainly possess that type of ceiling. S&P+ has five opponents ranked between #10 and #21 – Michigan, Stanford, USC, FSU, and Virginia Tech. Northwestern, Wake and Pitt then register as top-40 teams, followed by a still dangerous tier with Syracuse (71), Vanderbilt (75), and Navy (underrated at 85). S&P+ still likes the Irish a lot though, projecting 9.5 wins, which is the most common over/under win total I’ve seen.

#10 What’s your prediction, you coward?

The true sportsmath / blog boy answer here is to think about the season as a range of possible outcomes versus, and what’s most likely. I’m extremely high on the Irish defense – I think it’s likely to be a top-10 unit, with the possibility to be even better with young players breaking out. I trust the Elko to Lea transition to go somewhat seamlessly, and only the Stanford offense really scares me looking at the schedule.

The offense, however, has a far wider range of possible outcomes – anything from a top 50-60 performance as a floor up to a top-20 unit as a ceiling seem in play. In some ways I think this season could play out a lot like 2012, just with a less efficient offense, although close is possible. That run saw the Irish defense keep them in every regular season game, allowing them to survive sub-par offensive showings (the Irish scored 22 points or less in seven regular season games). All of the close games went Notre Dame’s way (5-0 in one-score games), but historically those contests should produce about an even won-loss percentage over the long run.

Optimally, like in 2012, the defensive strength can allow the offense to play fairly conservatively. It’s hard to feel good about the passing game until I see it consistently perform, and if the Irish can survive offensively if fairly one-dimensional. A best-case scenario is maybe the passing game (and as a result, offense as a whole) clicking in the right games and struggling in games the defense and talent edge are enough to still pull out wins.

If we consider a top 40 offensive performance a realistic middle-range outcome, let’s look at some comparable teams from 2017 teams that are fresh in our minds.

• Floor: ’17 Michigan (8-5, 27^th Overall in S&P+, #85 Offense / #10 Defense): The Wolverine defense was very good but had a couple of crippling flaws, including giving up too many big plays, and performed poorly in a few key games. The 85^th ranked offense may seem like an enormous drop-off for the Irish, but the rankings are so tight at that point that the difference in adjusting scoring average between #85 Michigan and #55 Utah is just 2 points per game.
• Ceiling: Rich man’s ’17 Wisconsin (13-1, 6^th Overall in S&P+, #41 Offense / #3 Defense) or ’17 Auburn against a non-insane schedule (10-4, 10^th overall in S&P+, #34 Offense / #5 Defense): Both teams were playoff contenders, that came with a conference title game of making it in. Strong defenses provided the backbone of each team, with offenses that were solid but not spectacular. Balanced attacks worked for the Badgers and Tigers, as both teams finished between in the top-25 of both Passing S&P+ and Rushing S&P+.

Put me down for something in the middle and ten wins. I think Michigan is the best team on the schedule, with the Wolverine defense by far the single best unit the Irish will face. It’s in week one, where maybe Shea Patterson and the offense are still a mess, and at home, but I’m not optimistic. As with the Georgia loss last season, I think ND will rebound loss well, including an extremely satisfying win over Stanford. I think Virginia Tech is a strong matchup for Notre Dame and a chance to make a statement on the road in a difficult environment.

Did anyone notice that in 2015 and 2017, the Brian Kelly didn’t lose a game to a mediocre or bad team (2016 never happened, don’t @ me)? I think that’s possible again with another team of similar caliber, with Wake Forest and Northwestern as the key watch-outs. The Wildcats in particular I think are the type of team that Irish fans should worry about the most – strong defense, experienced QB, and well-coached.

My best guess is that the second loss comes to FSU. The Noles lost a ton of production, but have elite athletes to challenge the Irish on both sides of the ball and will be playing better in Willie Taggart’s scheme in November than in September. A rebound win over USC, who have a lot of question marks and are more likely in my mind to be out of the top 25 at the end of the year than top 10, will provide some healing.

As always, the caveat here is that we know a whole lot but also nothing. Stay tuned for 1) an advanced stats preview for Michigan and 2) advanced stats reviews as the games begin. Some cool things are in the works (better data visualization! prettier charts and graphs!) so stay tuned and please share/comment/critique as you see fit.