With Michigan's non-conference schedule behind them I thought we'd look ahead to their next Big Ten opponent, and their interesting defensive system. Nebraska's new defensive coordinator Tony White runs a Rocky Long 3-3-5. We've come across this defense before—even installed a version of it 14 years ago. A different branch of it is the basis of the TCU and Ohio State defenses that felt frustrating to play against despite Michigan putting up loads of points in those games.

The Nebraska edition—the 3-3-5 stack—is about as pure of a 3-3-5 as you're going to find in today's modern, Everybody's Multiple age. And it's hard to argue with White's results with it thus far. Bill Connelly has them 28th in defensive SP+ this year, up from 61st in 2022. They're already up to 14 sacks after finishing with 21 and 20 the last two seasons, and are holding opponents to just 5.02 YPA (sacks included) passing and 3.32 YPC (sacks removed) on the ground. In their four games, the defense should have led to a victory at Minnesota, kept them in the Colorado game, and defeated a pair of G5 teams despite offensive struggles.

Where's this coming from? Everywhere.

[AFTER THE JUMP: Stack is back]

THE BRIEF HISTORY

According to Chris B. Brown of Smart Football, who wrote an article about this in our 2010 Hail to the Victors, the 3-3-5 Stack (or 30 Stack)'s origin story has two acts. That history is helpful in understanding how it works. The first was Joe Lee Dunn, who hybridized a 3-4 defense to frustrate the Run & Shoot offenses of his day at Memphis and later Mississippi State and Arkansas.

Dunn was a young DC at Memphis and was about to play USC without enough defensive linemen. The plan was actually more of a 4-3 philosophy—meaning everybody in the front has one gap to defend. But really he came to it by working backwards from his problem:

1. Replace a defensive lineman (expensive) with a safety (cheap, readily available).
2. Have a LB aggressively shooting the gap that the defensive lineman used to be in.
3. Have a DB aggressively replacing the spot where the linebacker used to be.

They're the same picture.

His pressures and run fits were the same, but what he lost in size Dunn won back in the fact that his linebacker-DL could appear in any gap, and this screwed up how blockers prefer to react, messed up run angles, and took away pre-snap reads, which are quite important to gap defenses that often decide where to attack based on the defense's alignment. If you're the right guard and have a defensive tackle lined up on your right, you think to set the protection or the downblocks rightward so you can block him. If you're that same guard looking at a 3-3-5 pre-snap alignment, how are you supposed to figure out that the Spur is replacing the SAM who is coming to your right?

Another thing it messed with, Dunn found, were the base assumptions that offenses were drilling into their quarterbacks' heads. Every play looked like a light box pre-snap, and like a weird blitz post-snap. If you were an offense that checked into run plays that attack open gaps you'd find yourself running into a hard node. If you were an offense that threw at the pressure, you'd be constantly throwing at zone defenders replacing the guy the pressure was coming from. It was doubly effective against the era's passing attacks. Run & Shoot offenses liked to use running backs as receivers to stretch coverage responsibilities. But faced with more complicated pressures, they found they needed to keep the RBs in the backfield as blocking wild cards. As offenses converted RBs to blockers, the linebackers who used to chase them into the flats could find other ways to be useful.

The second evolution was under Lou Holtz and Charlie Strong, who used it to turn around South Carolina. Holtz had a hard enough time getting Bob Davie to come run his 3-4 system at Notre Dame; when Holtz took the job in Columbia he hired former Irish DL coach Charlie Strong. They inherited a Gamecocks program that had won a single game in two seasons, and didn't have enough defensive linemen either. But what they did have was a big nose guard who could demand doubles. Strong took Davie's 3-4 philosophy, where the defensive line two-gaps their opposites to keep LBs clean to flow behind him, and incorporated it into the 3-3-5. He also added all kinds of stunts.

This amplified the number of ways the defense could attack. Offenses wouldn't know who's two-gapping or attacking any given gap. Power running teams found it especially frustrating because the gap they thought they were cracking open would often turn out to be filled with blitzballing linebackers. As Dunn developed his system he had incorporated hybrid players and different ways to take advantage of that hybridization, but philosophically he was always trying to get back to ways of running his system with lighter players. It was Strong who really leaned into a new, Stack-specific philosophy of "Anyone Could Be Anything." He would bring seven or just three, and kept meticulous notes on how defenses reacted to anything he did so he could find ways to "cheat." Dunn's defenses were always sound; Strong's defenses would leave you a badly defended gap or a cornerback in an impossible conflict in order to overload how he thought you were going to react. Good luck finding it.

As other teams adapted Strong's concepts they found different ways they could exploit this. For example Jeff Casteel, then Rich Rodriguez's DC at West Virginia, was the first to realize he could use a quicker, tiny MLB and have him line up further from the line of scrimmage. If it won't trigger you, here's a video by Casteel that explains how they ran it against different fronts.

Meanwhile Dunn had passed along his system to his former New Mexico DBs coach Rocky Long, who installed the system at Oregon State and UCLA in the 1990s before becoming New Mexico's HC himself for most of the aughts.

If Dunn is the 3-3-5's inventor and Strong was the guy who made it commercially viable, Long was the guy running the R&D lab for the next two decades at UNM and as Brady Hoke's DC then replacement at San Diego State. Long adjusted the system to RPOs by having the MLB drop back even further—almost a third safety—which he called The Aztec(<—Ian Boyd's must-read article on this from 2017). This was initially a way to play a Dantonio Quarters defense without putting his safeties in can't-win run/pass conflicts, but it adapted well to the 1-high looks that most 3-3-5 practitioners prefer.

With Long for all these stops were his lieutenants Danny Gonzalez and Tony White, until Gonzalez took the DC job at Arizona State under Herm Edwards and brought White with him. White became ASU's DC in 2019 then got the same job at Syracuse. When White went to Nebraska, his old mentor Rocky Long replaced him at Cuse.

PHILOSOPHY: BE HARD TO READ INSTEAD OF HARD TO MOVE

Here's Dunn describing the defense (via Jack Harris of Sundevil Source)

"We had undersized defensive linemen, so we converted our speed into the strength of our defense and pretty much our gameplan was to fire in all directions and then let nothing over the top in Cover 3. … People think of it as a defense to defend the spread attacks, but in reality it was a defense to attack pro-style offenses with undersized defensive players."

The thought process here is if Power run schemes are particularly lethal because they take advantage of the offense's advantages in strength and size, smaller defenses can make up the difference by using their speed to close down open space when it appears. DTs who can get blown off the ball by double-teams are more likely to be slanting away from those doubles (see the DE on the bottom hash in the gif above from Boyd's article), or else a linebacker buddy is about to show up next to them to make those singles, adding his movement towards the line of scrimmage to the force equation.

One of the reasons it's so effective is the blocking against this defense is so different than what linemen are used to. Those guys are used to seeing anyone on the line of scrimmage as an immediate threat, and anyone backed off as a problem to be dealt with as he comes. As a result the offense often expends too much effort moving out the few DL and are out of resources when the guys off the line suddenly engage.

Having watched Michigan run Duo for two years you see the nose getting comboed around and think the defense is in trouble as soon as the MLB picks a gap. The MLB has a gap, however. He's already heading towards it. The WLB isn't because he's got to watch the backside, but he's also unblocked so he has the leisure to wait out the handoff then clean up.

But it's the constant guesswork of who's going to appear where that really replaces raw physicality as the determiner of play results. Here's an example from the Colorado game where Nebraska put three down linemen on the line of scrimmage then saw their nose blown way off the ball by a double. Your brain interprets that as a play that should get 6-7 yards. It gets only a couple.

Why? I mean, they RPO'd the LB on the bottom, so there are just five guys for five blockers and a running back.

Except the DE on the top isn't coming where he's expected to be. By diving inside the B gap he's got the left tackle all discombobulated. The WLB is probably intending to scrape around the back of there, but because the LG is focused on him he's not turning around to stop the DE from getting into the backfield.

A moment later the double on the nose has blown that guy well out of the gap. But the MLB behind him is playing off of the nose's result. The NG doesn't have to hold up—just slide over a bit and keep the RG occupied while the MLB slips down into the run gap.

Even with the other DE (on the bottom) getting well kicked out and the nose now firmly inside the hash mark, the MLB has a pretty clear path to come down and thwack the RG and force this back to where the backside DE is coming through.

The RPO'd guy is never relevant. The nose and the bottom DE getting moved out by their blocks are never relevant. What's relevant is nobody blocked the DE on the top, and there's no way to escape him because the MLB could come down. The WLB is now on the top of the formation, protecting the backside edge from a bounce.

As for the linebackers hanging back, well, the OL are free to release on them, but fair warning: they're quick as cats, and as hard to corral in space. The advantage gained by having those LBs way off the line of scrimmage is they delay decisions. They're harder to pass against, but also harder to read for optioning purposes. Here the 3-3-5 swallows a zone read that's supposed to put the offense +1 in the box because Shedeur Sanders can't figure out if he's reading the WLB or the WDE.

The other thing about this defense is it's deceptively aggressive. They *look* passive because they're hanging off the line of scrimmage, dropping eight, etc. That's mostly a lie. Nebraska will stack the line of scrimmage and play Cover 1 or Cover 0 often. They will also threaten blitz then use their defensive backs to aggressively attack your quick outlets. Colorado got caught in an RPO here because they were certain this Spur is going to going attack off the edge, leaving the WR room underneath a CB who's got to stay high.

Then everybody but three rushers drops back, which hard enough to throw into. But the real problem for Colorado here is the cornerback isn't bailing deep like he should if it's Cover 3.

And since this was an RPO, the blockers are in no position to deal with the DE on the top, who gets a free sack on a three-man pass rush.

Disguising coverages isn't the main reason this works. The Spur wasn't deep enough to pick it off until late in the progression. The problem Sanders ID'd was the cornerback didn't even drop to the sticks and then started driving, so any throw completed is going to be immediately swallowed for no gain. The CB can play that aggressively despite help from the Spur down in the flat because he knows this play isn't going to last long. They know they have the QB thinking about quick pressure and finding the hole in the defense, not dropping back and surveying.

I figure you've got a good enough memory of the Fiesta Bowl that I don't need to show you how aggressively TCU was flinging linebackers into gaps despite showing a passive 3-3-5 front before the snap. Here's a pleasant memory that makes the point. Note that the safeties are also stepping down to provide a second wave, essentially playing linebacker. The safety's tackle attempt on Edwards here was at about the 28 yard line; had he connected it's a 3-yard gain.

HOW DO YOU BEAT THIS STUFF?

The 3-3-5 is designed to force offenses into low-success-rate football. Offenses predicated on moving the ball consistently find this infuriating because it forces them to go hunting for big plays. The tendency—and I fear this is what Harbaugh wants to do—is to get "Big", bringing in more linemen, tight ends, and fullbacks to increase the gaps the offense can threaten. The thinking goes if you complicate everything for the defense they're going to have to either commit more linebackers and safeties to gaps they're not big/practiced enough to fill.

The problem with that approach is it just gets the 3-3-5 to act even weirder. They thrive on hybrid players flying to the ball, so if you allow them to add more guys to the box you increase the ways they can confound you. If an opponent is determined to play rock-paper-scissors, you have to do so as well.

The trick, if you ask me, to beating these defenses is to punish their aggression with big plays. Remember this DE who slid inside the left tackle?

Well if they want to spill while the help is engaged inside, run out the backside,

…or just run it outside (though Nebraska isn't TCU in this regard; they stopped a 4th & 3 jet sweep in this game).

If they're showing blitz then bringing three, protect, read, and wait for one of their hybrids to lose one of your specialists.

If they're hammering down on quick throws like this CB…

Put the quick throw downfield.

Colorado had a tunnel screen that should have been a touchdown when it caught Nebraska blitzing…

And Michigan put one of these on tape against Rutgers despite not running one since 2015 according to our charting. That's a weird thing to prepare for Rutgers, which plays a lot more man defense than Michigan's other opponents in September. But if you think back to 2016, Luke Fickell's response to Don Brown's five-man pressures was to run tunnel screens from all kinds of looks, forcing the defense to leave guys in the middle of the field they'd rather be blitzing.

It's a fair bet that Nebraska's new defense is preparing to frustrate the hell out of Michigan at home. Most likely they're burning all kinds of practice time right now on ways to confound the Wolverines' multi-TE sets. Hopefully Michigan is practicing ways to maximize the damage when they win their guessing and execution games. Given Nebraska's offense is leagues behind their defensive renaissance, one good half should do it.