clock menu more-arrow no yes mobile

Filed under:

Duck Tape: Film Analysis of Michigan State

A preview of Oregon’s bowl opponent in Santa Clara

Oregon v Michigan State Photo by Streeter Lecka/Getty Images

This piece is a complement to the personnel breakdown of Michigan St that went up on Friday morning. That article went through the players and basic gameplans, this one will focus on tendencies, vulnerabilities, and film study. Stats in this article, as in all my film study, refer to snaps prior to garbage time and exclude trick plays and obviously broken ones, and I label plays as run/pass/screen based on the intent of the playcall.


This is pretty clearly a run-first offense, rushing 60% of the time on 1st downs and 75% on 2nd & 6 or shorter. The problem is that Michigan St just isn’t very good at it, coming in under 48% at per-play rushing success. Part of this is some issues at back, since they were missing their starter for much of the year and the third-string guy in their two-back system wasn’t getting the same yards after contact that the first two would. But I only saw a small handful of plays where the backs had good blocks to work with and made some mistake of their own … mostly the problem was that the offensive line, tight ends, and fullbacks just can’t run block very well (reminder - you can right-click or long-press any of these videos to watch them in 12 or 14 speed):

On the first play, the #97 S&P+ rushing defense has a light box against two tight ends and a fake sweep on this inside zone rush, and still disrupts these blocks effectively. On the second play (at :07), MSU hands off on the sweep and gets some decent outside blocking by the skill players, but there’s just not enough of them in this play design, plus the center misses his second level block on the MIKE, meaning the defense has six defenders against four blockers and that’s more than enough to shut this down. On the last play (at :21), the offensive line is destroyed in a comical manner.

MSU dropped back to pass over 100 times more often than they rushed, and two factors stink of poor run-pass optimization here: first, a heavy tendency to pass on 2nd & long and all 3rd downs (a 3:1 ratio, including 17:1 on 3rd & 6 plus), despite rushing on such snaps finding success nearly 60% the time vs only 35% success passing; second, by far their most effective option on 2nd & short is taking deep passing shots, and yet they only did so on 12% of those plays. But regardless, they’re not really an effective passing offense either:

The first play is fairly typical of pocket breakdowns (which happened on about a quarter of all dropbacks) - the slot has beat his man and this might have been a touchdown with just a second more protection or just being able to step up in the pocket, but the right tackle is completely beat and really the entire line is collapsing, so the QB doesn’t have anywhere to go even if he had sensed the pressure. On the second play (at :13), MSU has kept home eight blockers against just a four-man rush and the QB is still flushed - because they can only send two men downfield and neither gets open, and the QB has a case of happy feet. The last play (at :29) shows a true freshman LB playing man coverage against a WR who has more than enough speed to burn him - the QB should just throw it deep and let him run under it, but as I observed on about 40% of all deep and intermediate passing plays, the QB is making an unforced error.

With a season’s worth of data I should have been able to make a video about MSU’s screen game, but shockingly for a team with a shaky o-line and a QB in need of confidence, they threw a screen on fewer than 3.5% of all snaps (in three games they attempted only a single screen pass, and against Michigan they threw none at all). They’re underwater on per-play effectiveness on those, though that’s true on pretty much every play they run, mostly because of poor blocking but also a few wobbly throws from the QB.


As readers have no doubt heard, MSU has the #1 rush defense in the country. That’s partly a self-fulfilling prophecy - teams just quit trying to run against them. But it’s mostly because of a pretty phenomenal set of defensive line starters:

The first play is the quintessential MSU rush defense - the DTs instantly collapse the line despite taking combos, the MIKE crashes the instant he reads run, and the DEs and remaining backers play clean-up. The second play (at :06) is such a beautiful illustration of an arm-over move by the DT I had to include it. On the last play (at :21), their best DE single-handedly blows up this play (something I recorded him doing over 100 times on the season) by putting that poor tight end on skates.

There are some vulnerabilities I identified in the rush defense, however - they involve taking advantage of their overaggression and a surprising lack of depth at DT. Interestingly, the type of run play that they defend least effectively — by a large margin — is fairly simple inside power, because if you don’t let their d-line bully you then you’ve got a pretty good shot. Some examples:

On the first play, MSU has their backup DTs in (weirdly, they appear to only have two and they always play together, never mixed with a starter), and as you can see they’re much less effective at absorbing blocks. The crossing double-A gap blitz gets picked up and the back runs over the safety - the dirty little secret of MSU’s defense is that past the dominant starting DTs they have a tough time defending the middle of the field. On the second play (at :06), again the backup DTs are easily cleared, the backers are washed outside, and if the back hadn’t lost his feet he probably could have put a move on the safety and ran the distance. On the last play (at :20), MSU has gone to its interesting 2-4-5 package it uses on a lot of 3rd downs - they’re expecting a pass and when the blitzer misses the handoff there’s no one left in the middle for this surprise run.

Pass defense is only a little less effective, still well above water on a per-play basis, but harder to assess the moving parts. The starting defensive line contributes to most of the success, and while I think the corners are pretty good, I can’t really be positive because a) most opponents’ passing plays don’t have long enough to develop, b) broadcast camera angles make judging CB play really tough, and c) injuries to a couple key players plus their best guy sitting it out for the draft mean I don’t have a ton of data on the two guys I expect to start in the bowl. Here’s some representative plays:

On the first play a nice four-man pass rush gets home, affects the throw, and results in a pick - note the great cover-1 defense across five receivers and the free safety playing the ball. On the second play (at :31), there’s no real pass rush (note the backup DTs) and a pretty wonky zone handoff, but this is one of the few plays where I got to watch their starting CB from last year (who’s returned from injury to play in the bowl), and he makes a great hit on a big TE to knock out the ball. On the last play (at :46), we’re seeing their nickel package again, this time working as it’s supposed to with the stand-up OLB/DE coming off the edge hard, flushing the QB, and picking up a holding flag - and good thing too, because as is typical of this package the back seven don’t actually do a great job in coverage if the QB can survive that edge rush.

Again, I think there are some identifiable vulnerabilities here. For one, they almost comically refuse to defend the six-yard stop route in zone coverage - it’s not schematically interesting so I’ll just describe it, they surround the TE with two backers and a safety even if he’s standing on the line to gain and don’t move to tackle him until it’s caught. For another, they really don’t handle hurry-up snaps very well, as I’ve got multiple substitution infractions and LBs well out of position waiting for the play call when the opponent snaps it quickly (of course, they really only saw quick snaps from Utah St and the former Oregon OC at Nebraska). Also, the safeties don’t have the speed to handle quick routes on their own; they need underneath help and the backers are too aggressive for it. In general I find this pass defense to be fairly predictable, with easily manipulable keys. Some examples:

On the first play, watch the MIKE closely - he steps up into a run read for no discernible reason, and realizes too late he’s abandoned his underneath duties on the slant route, which the safety can’t possibly cover on his own. On the second play (at :07), with their best DE on the field side the QB can roll away from the other one and has an easy high-low read after the backers are frozen on the play-action fake - he hits the wide open drag route but the flat-footed safety has left open the X-receiver on the post for what might have been an 87-yd touchdown had the QB seen him. On the last play (at :23), once again we’ve got that nickel look, and even with a clever disguised blitz, having their backup DT try to run a loop and their best rushing OLB/DE back out into coverage means the QB has plenty of time to throw, and look how soft the zone coverage is of the underneath hole (weird to call it that when it’s 20 yards downfield) … again, if you can survive the initial onslaught of this package, you can pick up pretty big gains against it.

Finally, by a pretty wide margin the types of plays MSU is worst at defending are outside RB, flanker, and split screens. This is largely because they use the backers’ keys against them and play into the vulnerabilities with their fast but undersized STAR and his twin brother at safety:

The first play is a mixed bag - the backers are sucked in hard by the play action fake, leaving the DBs with one-on-one blocks on the outside, and neither exactly gets off it, although they do delay the receiver from turning the corner - fortunately the other safety is fast enough to make it over in time for a sure tackle. The second play (at :08) is better executed by the offense - the MIKE checks out of his blitz but he can’t do anything about the RB sprinting out for the screen, and again the DBs can’t get off their blocks fast enough. On the last play (at :16), I’m not sure how much more blatantly the offense could be telegraphing screen on the first second after the snap, but inexplicably the backers are surprised by it and have flowed away, and the safeties are similarly slow in recognition to come down - instead of the starting backers or safeties, someone else ahead of them seems to know the way.