After finishing dead last in offensive DVOA last season, the Steelers offense needed a jolt. The once-prolific offense that ranked 6th in the NFL in 2018 was suddenly the team’s glaring weakness. With Antonio Brown’s menacing isolation routes no longer a part of the offense, the remaining skill players failed to achieve anywhere close to his production in the passing game. It didn’t help that starting quarterback Ben Roethlisberger missed most of the season with an injury. Like a domino effect, the rushing offense plunged too, finishing 30th in run DVOA.
While Brown’s release and Roethlisberger’s injury were daunting obstacles, there were tactics that offensive coordinator Randy Fichtner could have still turned to. For example, the Steelers’ used play-action at a dead-last 12% of passes in 2018, and repeated that clip in 2019, according to Steelers Depot.
This is malpractice. Play-action passing, as Robert Mays once wrote, is the NFL’s version of the corner 3. It’s an area where defenses are highly vulnerable, but haven’t been exploited sufficiently. According to Football Outsiders’ Scott Spratt, the league-wide DVOA differential between play-action and non-play action passes was between +11.3% in 2018 (no data on 2019 yet). For perspective, that gap is about the same as the one between the 7th-ranked 49ers offense and the 20th-ranked Browns offense.
The Steelers can look within the division to identify another of their major offensive inefficiencies. The Baltimore Ravens’ prolific offense last year defied all kinds of norms; among them, the sky-high 34% rate of snaps with a man in motion. No other team did this more than 20% of snaps, and only 10 did so more than 10%. The NFC Champion 49ers and the Super Bowl Champion Chiefs both ranked among them. The Steelers did not.
According to ESPN, there was a significant difference in expected points added per play (EPA/P) between plays with motion at the snap and plays without it. The difference was 0.11 EPA/p on run plays and 0.08 EPA/p for pass plays. This is especially important in the run game because all but 4 teams averaged less than 0.01 EPA on all rushing plays last year and all but 5 averaged negative EPA. If you don’t have a gifted runner like Lamar Jackson or Kyler Murray at quarterback (Ravens and Cardinals) or historically talented run blockers (Colts and Cowboys), there is no evidence that “establishing the run” improves your probability of winning NFL games. That is, unless your run schemes add value, and motion at the snap does that.
The Steelers’ deficiencies in play-action and lack of pre-snap creativity make their hire of Matt Canada as the quarterbacks coach extremely curious. A colorful illustration of why I say this comes from Lane Kiffin. “Watching LSU’s offense [in 2017],” Kiffin said. “I left Orgeron a message, ‘Hey, did you have a bonus in the contract for your offensive coordinator, for every shift he got extra money?
Canada was a college offensive coordinator every year between 2007 and 2018. In his stints at Indiana (2007-10), Northern Illinois (2011), Wisconsin (2012), North Carolina State (2013-15), Pittsburgh (2016), LSU (2017), and Maryland (2018), Canada brought a novel set of shifts, motions, jet sweeps, shovel passes, and play-action passes.
That novelty became perceived as gimmickry when his offenses didn’t live up to standards, as it didn’t at LSU. But when Canada’s offense clicks, it is a force of devastation. His greatest feats were racking up 43 points in a victory over the eventual 2016 national champion Clemson Tigers and 51 points on the 2018 Big Ten champion Ohio State Buckeyes. Teams with Nate Peterman and Tyrrell Pigrome at quarterback laid waste to a bounty of future NFL defensive players. Those heights are why Canada’s arrival in the NFL fascinates me as an under-the-radar offseason story.
The Easiest Offense in America?
“Our offense is the easiest in America,” Canada once said at a high school coaches clinic in 2017.
“No way! This is it?” he later told the crowd smiling. “This is it!”
Canada was being facetious about the totality of his offense, but he’s right about the simplicity of his play-designs—if not the pre-snap action leading up to it. The bread-and-butter play of his offense is hardly exotic. It’s the Inside Zone run scheme nearly every NFL offense since the Joe Gibbs-era Washington Redskins has installed.
His other signature play, the jet sweep, has been around since the 1940s when Gene Beck won California high school championships with it as its centerpiece. In the 1970s, other high school coaches got their hands on the reels of Beck’s offense out of old plastic storage bins. One of them, Mark Speckman, became the Willamette offensive coordinator (NCAA Division III) and introduced the jet (or fly) to Canada when he was at Indiana.
These two plays are effective in Canada’s offenses because he packages them together. His offenses run the Jet going one direction and Inside Zone in the opposite direction. Most of his blockers step to the side of the Inside Zone action, while the remaining few step to the side of the Jet action.
The rule generally seems to be that the strong-side offensive tackle and everyone else over to the direction of the zone side is a zone blocker. The in-line tight end, and sometimes an H-back (or fullback) too, will be the lead blocker for the jet sweep side. Like the read option, this Inside Zone/Jet Sweep pairing puts a defender—usually the edge rusher—in conflict. This is why it is more efficient than standard zone schemes.
The variety kicks in with how many players Canada gets involved as the runner on jet sweeps or end-arounds. Sometimes, the H-back/fullback will get the end-around from motion. A running back might start in the slot and get the jet sweep. Even offensive tackles have gotten the ball as the motion man. It’s all the same basic play-design though.
On short-yardage and red zone situations, Canada’s calling card is a true option play with a pulling guard. This Power Read Shovel, or Power Shovel Option, play again typically forces an edge rusher in conflict because the quarterback is reading him. If the EDGE crashes to stop the Power play, the quarterback flips the ball to a player (usually the H-back) on a shovel pass. If the EDGE stays put in his gap, the running back gets the handoff and trots through the B-gap.
There are again other variants of this basic play. Sometimes the running back aligns on the play-side to receive a toss instead of a standard handoff from the back-side. In that look, a triple option can be incorporated, with the quarterback able to keep for a run. This was a play that Urban Meyer ran with the University of Utah back in 2004.
Instead of a give to the running back, the alternative option to the shovel can be a passing play too. It can be a snag concept, a stick concept, a vertical-out route combination, or a rub route to the running back.
Half-field reads are typical as true passing play-calls, not just in Canada’s option game. Those classic West Coast offense passing plays—Snag and Y-Stick in particular—are counted on regularly. The bootleg is also common as yet another stress on edge defenders who have to account for the play-fake and quarterback keep on opposite sides. Like those West Coast three-man route combinations, boot concepts cut the field in half for the quarterback and make his progression simpler.
This is why the Bootleg Sprint-Out meshes well with his run schemes. The play-fake to the running back functions the same way a jet sweep does. It freezes the unblocked edge defender and gives the quarterback time to roll out on the bootleg and scan the field. The route combination of a corner by the X-receiver and a quick out by the Z produce a Hi-Lo read. The play-fake usually occupies underneath zone defenders long enough to get one of them open. And if the defense is in man, that’s an invitation to clear out the defensive backs with vertical routes and use the jet sweep to exploit open space.
In fact, most of Canada’s play-action passing plays consist of half-field reads. Bootlegs generally stress defensive ends as the offensive line simulates zone run plays, while standard five or seven-step drop play-action passes typically target linebackers who are conflicted by their run and pass responsibilities. Those latter plays work best when Canada’s teams pull the back-side guard to simulate the Power O gap scheme.
Canada is not the first to use this play-action protection scheme at a high clip; the Patriots have done so throughout the Bill Belichick era. Chris B. Brown wrote about how, at the college level, Andrew Luck (Stanford) and Robert Griffin III (Baylor) both frequently connected on deep posts and deep crossers with a pulling guard in front of them. Both schools leaned on Power O in their run games too, with Stanford doing so in the I-Formation and Baylor using a Power Read/Inverted Veer option play from Spread formations.
Even though Canada uses Power O sparingly as a run play (and when he does, it’s usually paired with the Shovel Option), this particular play-fake scheme sets up his most explosive plays too. This might seem counterintuitive, but it’s in line with Ben Baldwin’s findings that play-action success and rushing success have essentially no correlation. The easiest offense in America? Maybe not. A simple one that puts defenders in conflict almost every snap? Absolutely.
The Read Option Without the Read Option: A Conservative Run Game for the Analytics Age
When we’re thinking about the potential for Canada’s ideas, the historic Seahawks’ Legion of Boom defense matters a lot. Pete Carroll’s teams back then played Cover 3 and Cover 1 virtually the whole game. Strong safety Kam Chancellor was basically a fourth linebacker, giving Seattle eight defenders in the box almost every snap.
This is the blueprint that almost all NFL defenses are now following, with the obvious flaw that none of them have Chancellor and only the Ravens now have (older) deep-middle safety Earl Thomas. Sports Info Solutions says that all but 5 teams played single-high (or middle-of-field-closed) coverages the majority of snaps in the 2018 season.
We know that box count and field position explain 96% of rushing-yards-per carry totals, based on Josh Hermsmeyer’s findings. This means offenses average less yards per carry against 8 men in the box than when they face 7, and less yards per carry against 7 men in the box than 6.
One solution that Spread offenses have deployed for years is to use 11 personnel. After all, Hermsmeyer shows the number of rushing plays with 11 personnel and rushing plays with 6 or fewer defenders in the box have a strong positive correlation. Another is to use the quarterback as a runner on option plays or designed QB runs. Lamar Jackson and Kyler Murray were the most efficient runners last season in terms of EPA/p or success rate by a wide margin. Most teams are reluctant to expose quarterbacks to contact though, in spite of Warren Sharp’s findings that designed QB runs have lower injury rates than dropbacks.
Another solution is what Canada brings: a wide array of motions and shifts that effectively reduce box count at the snap. Instead of freezing a defensive end or linebacker by having your quarterback read him, freeze him by making him think either the running back or a receiver in motion could be the ball-carrier. It’s basically the effect of the read option without the read option. Below is an example of that tactic.
In the third quarter of the 2012 Big Ten Championship, Canada gets Wisconsin’s offense into a Tight-Wing formation (two tight ends on the same side). It draws 8 Nebraska defenders into the box. This is the way then-head coach Bo Pelini—a former Carroll assistant—often responds. But when Melvin Gordon (no. 25, initially aligned as the left wide receiver) executes jet motion, that 8 man box becomes a 6 man box against the Inside Zone. The arc blocks by the two tight ends lure 2 defenders and the deep safety away from the action.
Hermsmeyer’s data shows that the number of defenders in the box is more predictive of yards per carry than the relative blocking advantage. This is why five Wisconsin blockers are sufficient against the remaining six Nebraska defenders. The jet play-fake freezes the defensive end and the offensive line forms a perfect wall on the other five Cornhuskers. All Montee Ball has to do is break two tackles.
Earlier in the game, the two tight ends set up the same play-design for a big gain from Gordon on the end-around. It begins with one of Canada’s most common pre-snap shifts, Double TE trade, in which he moves two tight ends from one side of the formation to the other. Nebraska is in a 4-4 Over front, with a box safety as the overhang defender on the edge. The key to this play is how the SAM linebacker (no. 54) chases after the Inside Zone action. This leaves seven of the eight box defenders chasing ghosts.
Nebraska actually had a good coverage adjustment. Its initial coverage was Cover 3, but the boundary cornerback rotates to the deep-middle when he sees the jet action. The deep-safety is watching Gordon too and he rolls into the box. Now he has a better angle to attack Gordon than the boundary corner would have, and the overhang defender can occupy the lead-blocking wing tight end.
However, the analytics on defender count versus relative advantage comes into play again. Explosive plays are more likely if the runner has space, even if he doesn’t have as many blockers. When an athlete like Gordon gets the ball, he can exploit that space.
A similar combination of pre-snap action sets up Pitt’s first touchdown against Clemson in 2016. This time, a trade by H-back George Aston from the right side of an empty set to the formation’s left wing precedes jet motion from left slot receiver James Conner. Aston gets shadowed, signaling man coverage to quarterback Nate Peterman. With no deep safety, Clemson prepares to run a Cover Zero blitz.
The Power Read Shovel comes in a unique form here with the jet motion man as the give option. Unlike Canada’s jet sweeps and zone runs, this one truly is an option play. But once again, the quarterback keeper is not one of those options.
Right after the snap, three Clemson defenders in the box are unblocked. This is by design, as five defenders are easier to block than eight. When Peterman sees the defensive end and edge blitzer both charge Conner, he flips the ball to Aston. Meanwhile, another edge blitzer runs toward air on the other side. The play-side tackle and pulling back-side guard each smother the two linebackers. With the other offensive linemen executing their down blocks on three Clemson defensive linemen, Aston has a clear lane.
Canada didn’t exactly blindside Clemson by relying on the Power Read Shovel in 2016. He used it the season before at N.C. State and it yielded a Jacoby Brissett rushing touchdown. The version he called that time was a triple-option, with a quarterback keeper added to the toss and shovel options. The toss was especially effective at neutralizing the edge blitz. The pulling guard took out the defensive end and freed up the tight end for an arc block at a favorable angle.
Besides the shovel option, Canada only sparingly used gap-blocking schemes in conjunction with the jet sweep action. A Buck Sweep play by LSU against Mississippi State in 2017 showed the potential of this combination. The TE Trade leaves the run gap (the left C-gap) defended by a safety (no. 4). The right linebacker (no. 10) rotates from off the ball to the edge in response to the trade, while the safety replaces him at the second-level. The down-blocks by the tight end (no. 84) and left guard (no. 77) free up the left tackle (no. 63) to trap the defensive end (no. 10) and the center to wrap up to the safety.
Remember what I said before about 11 personnel and its boost to run game efficiency? This play is one where we see it. On the backside, we see a motion called jet twirl—jet motion one direction, followed by a pivot to jet motion in the opposite direction. Jet twirl usually comes from a slot receiver because the outside receivers are typically more effective blockers. Finally, the jet action freezes defensive tackle and future first round pick Jeffery Simmons (no. 94), so LSU does not even block him.
The Advocate reported that Canada visited with Spread-to-Run savant Chip Kelly and Pistol inventor Chris Ault in the late 2000s. It’s easy to see how their ideas allowed him to contain especially disruptive defenders like Simmons without even blocking him. At Oregon, Kelly’s Inside Zone schemes were paired with the quarterback ‘reading’ a backside defender, similar to how Canada’s jet sweeps often freeze that backside defender. At Nevada, Ault used the read option from the Pistol with quarterback Colin Kaepernick to stress a front-side defender, much like the Power Read Shovel Option does. Canada’s rushing attack might not involve the quarterback much, but it’s clearly rooted in the evolution of the Spread as much as ones that do.
The Power of Play-Action: An Intro to Canada’s Vertical Passing Game
Before Kelly and Ault’s ideas became part of his offense, Canada was better known for his passing schemes. A long-time Division III and high school coach, Mark Speckman, introduced the jet sweep package (or the fly, in his preferred verbiage) to him, in return for advice on passing concepts. In fact, Canada’s original scheme influence was Terry Hoeppner, his boss at Indiana who previously used a pass-heavy Spread system at Miami (Ohio) that helped Ben Roethlisberger get drafted in the first round.
Yet, Canada’s passing game became much more explosive with the jet sweep and other Spread ideas incorporated. His 2016 Pitt offense was the most successful one at pushing the ball down the field. The Panthers ranked sixth in yards per completion in the whole FBS; only the nearly-run-exclusive Flexbone teams (the three military academies and Georgia Tech) and Penn State ranked higher.
That team wasn’t an outlier for Canada. The 2011 Northern Illinois offense ranked 12th in yards per completion, the 2012 Wisconsin offense ranked 23rd, the maligned 2017 LSU offense still ranked 16th, and the 2018 Maryland offense ranked 22nd. Those passing offenses varied in efficiency, but when they did connect, they were all explosive.
The jet twirl motion shows BYU is in zone coverage; it turns out to be Cover 3 Buzz. The field safety rolls into the box to deal with the threat of a jet sweep or zone run. The boundary safety, meanwhile, takes steps to the middle of the field which leaves the cornerback on his own to deal with current Jaguars starting receiver D.J. Chark on the post. We see how trailing a receiver who can run a 4.34 40 goes.
That example used a zone play-fake, but as I mentioned earlier, Canada uses Power play-action more frequently and successfully. To linebackers who are taught to read their keys (the movement of the offensive linemen being an important one), a pulling guard is an effective diversion regardless of context.
The Post-Wheel play-design that Canada used against Clemson is one that disguises itself as an offensive staple. After gashing the Tigers’ defense with Power Read Shovel, the pulling left guard and wing tight end make the defense key onto this option play again. Clemson was already stretched thin to the boundary because of their Quarters assignments. Against a Trips set, Quarters defenses generally position both safeties away from the isolation side.
The main vulnerability here is that the isolation-side cornerback gets a “MEG” (man everywhere he goes) assignment, as Clemson decides to blitz one of its linebackers. The shovel play-fake works well here precisely because there is only one linebacker remaining in coverage. That linebacker locks onto the wing tight end, leaving running back James Conner wide open. The boundary safety technically is the one responsible for Conner, but there was no chance he could navigate from the trips side to the sideline in time.
A version of the Post-Wheel with Power O play-action and jet motion Canada used against Bowling Green in 2011 achieved the same result. Bowling Green went with a Rip/Liz call where one of the safeties slides into the box right before the snap. The pulling guard and jet play-fake leave the four underneath zone defenders charging into their run fits.
The box safety may have assumed the cornerback would handle the running back. The cornerback’s assignment in this pattern-match coverage is to play the deep post man-to-man once the stem reached the 5-to-7 yard threshold. This lures the corner away from the sideline. The deep out by the slot receiver on the other side, meanwhile, manipulates the deep safety away from the middle of the field. This means that the running back has no one covering him deep, setting up a long touchdown.
Against Virginia Tech, the pulling guard worked in conjunction with Pitt fullback George Aston’s release to the opposite-side’s flat. It created a throwing lane so wide you could drive a bulldozer through it. The pre-snap motion signals that the Hokies are in some kind of zone coverage—look at the cornerback remaining in the boundary with no eligible receivers to cover. Post-snap, it looks like this is a pattern-match coverage, something like Rip/Liz.
Aston’s release is crucial because it lures the linebacker away from the middle of the field. Downfield, the Panthers are executing a Mills concept, a vertical two-route combination that involves an in-breaking route by slot receiver Jester Weah. This concept is great against pattern-match coverages because it’s impossible to provide help to both routes. From the power play-fake to the route combo, Nate Peterman is handed an easy throw.
A natural counterpart to Mills is the Dagger concept (in route by the outside receiver, go route by the slot) which Canada deployed against Miami in 2016. The power play-fake directly contributes not just to the protection scheme, but the actual throwing window here. Peterman waits patiently for the middle linebacker specifically to chase the pulling guard, only to get stood up by Conner. Once that happens, that dig route is open.
The power play-fake also set up a throw directly behind the linebackers against North Carolina. The short motion indicates either man or quarters coverage. The route combination looks like an over paired with a shallow cross, similar to the popular Yankee Concept which pairs a post with a cross. Shallow crossers are brutal to defend from man coverage and even more when linebackers charge into gaps because of a play-fake. Judging by how the cornerback momentarily releases the crossing-route receiver and shifts his eyes to the running back, this looks like Quarters. That momentary hesitation, along with the linebackers biting on the fake, is all Pitt needs.
A favorite of Kyle Shanahan, Sean McVay, and Lincoln Riley, Leak (or Throwback) is a devastating play-action pass concept that I only saw Canada use at North Carolina State. The idea in plays like the one above is to roll the quarterback on a bootleg to one side, and wait for the defense to leave a receiver who slipped across the opposite side undetected wide open.
The pre-snap jet motion to the right toys with the Clemson linebacker (no. 44), followed by a play-fake to the left, then followed by the bootleg right with the Over route and pulling guard. All of this horizontal stress makes it easy for wing tight end Jaylen Samuels (no. 28) to release unscathed to the left after his chip block. The Steelers have used Samuels primarily as a traditional running back, but I think Canada’s prior usage of him as a “slash” player here, especially at the Wing, is his best role.
Like Canada’s other play-action concepts, Leak works well with a variety of pre-snap motions and formations. Against Florida State, a short motion by the no. 2 receiver into a Wing tight end alignment made this play hold up. The Seminoles rushed five, but the tight end chipped the defensive end on the outside zone play-fake to buy quarterback Jacoby Brissett (no. 12) time. With several future NFL cornerbacks, Florida State liked to play man coverage against N.C. State’s receivers. However, this cleared out the sideline for the tight end on the Leak route left by the time Brissett finished his Bootleg to the right.
Brissett is a capable NFL backup or low-end starter, but the beauty of these play-designs is you don’t have to even have a quarterback of that caliber when everyone else executes. As Brown says, “the guy to thank for it is that big pulling guard.”
The Money Play: The Flood Concept
Almost every NFL offense uses the Flood concept today because its route combination stretches the defense in three levels on just one half of the field. A quick out or flat route is typically the checkdown, an intermediate out is the primary read, and a go route is available if the quarterback wants to pick up chunk yardage. This is Canada’s most successful passing concept and it’s crucial to understanding the overall strategy of his offense.
A series of shifts against Clemson put the defense in conflict on a version of Flood utilizing a bootleg. Before the snap, Clemson’s defensive line reacted to the shifts by moving over one gap away from the play-side. The left defensive end’s closer alignment to the center gave Peterman more room to avoid the rush while on the bootleg.
Clemson appears to be in Quarters coverage. This means the intermediate out by no. 6 will make the safety “match” him. The “match” assignments on the running back are what create the conflict. In Venables’ Quarters scheme, both linebackers eye the running back. The one whose side the running back flows to matches him, and the other becomes a robber or ‘rat’ defender who provides help. This is a problem because the half-field read is happening in the exact opposite direction of where the ‘rat’ linebacker flowed to. With the rat nowhere around to help, Peterman waits patiently for the ‘matching’ linebacker to abandon his assignment.
The half-field read that Flood provides also forces defenses to waste some of their defenders when in zone coverage. The bind is that, if you don’t play zone against Canada, you’re now vulnerable to the jet sweep. As a Penn State fan, I’m unfortunately aware of the onslaught that Canada’s Pitt offense dealt Brent Pry’s zone and quarters coverage-centric defense in 2016.
Against the Flood, Penn State had a curl-flat defender drop out of the box and a deep-third defender in the boundary—neither of whom was anywhere near the play-side. The two hook defenders both got sucked in by the Power run play-fake and were nowhere near the play either. On the play-side, the outside cornerback drops into his deep-third zone and leaves the intermediate part of the field wide open for the slot receiver.
Canada’s Maryland offense used Flood to beat the Buckeyes’ man coverage by aligning all five eligible receivers to one side. The Buckeyes leave a safety on the weak-side, likely in fear of the jet sweep or the inside zone coming that direction. With the running back and tight end staying in to protect, that takes two more defenders away from the action. At that point, it’s up to the slot receiver who’s running the vertical route to win leverage off a double move.
The pre-snap slot jet motion signals Texas is either in zone or a pattern-match coverage. With Texas in quarters coverage, that meant the boundary-side linebacker had to carry the slot’s post route up to the safety. This frees up the boundary receiver who gains inside leverage on the cornerback who realizes only too late that his inside help to defend the crosser is gone.
The same route combination yielded a huge gain against Ohio State on the go route. The tight end trade twirl before the snap again signals the defense is either in zone or a pattern-match coverage. It turns out to be Quarters again, and both safeties lock onto the post route. This is because the design of this play left the left safety with no good options. If the left safety released the post route to help on the go route, he risks leaving the post receiver free to use a double move to get himself wide open down the sideline. Instead, the go-route receiver gets the one-and-one and simply outruns the cornerback.
Against Texas A&M, Canada’s LSU offense created the illusion of a Flood concept. This is an Inside Zone play-fake, followed by the fake Flood, followed by a screen back to the running back. The shift from a Wing-Twin formation to an Offset I featured no defenders shadowing the skill players, signaling that the Aggies were not in man coverage.
In single-high coverages (that aren’t man), the deep safety is supposed to cover the no. 3 receiver (the in-line tight end). However, the MIKE first has to carry the no. 3 temporarily because the safety cannot cover underneath routes. With the rest of the defense pattern-matching, no one has their eyes on the running back. Not until the MIKE turns around. By that time, the center mauls him and the cornerback simultaneously.
The whole point of jet motion is to stretch the defense horizontally, so when a three-level stretch like Flood is paired with it, it’s a nightmare to respond to it. If you live in zone coverage, you might contain the run, but you’re only built to cover two levels, not three. If you play man-to-man, the flats are wide open for a big run gain. Damned if you do, damned if you don’t.
The Defense Rises: The tactics that work best against Canada’s ideas
The fascination I have with Canada’s offense also stems from a curiosity about how defensive coordinators will respond to his ideas. The teams that did the best at stifling Canada’s offenses were the ones that defied predictability and engaged in deception of their own.
My Nittany Lions were one when they redeemed themselves with a 38-3 win in 2018, as Canada’s (admittedly less talented) Maryland offense entered State College. Defensive coordinator Brent Pry approached this matchup very differently than the way he did in 2016 against Canada’s Pitt offense.
The infamous 99 yard opening drive in 2016 featured Penn State deploying eight defenders in the box and only one high safety every play. Pry’s strategy in 2018 was to use his more usual two-high safety looks from the beginning. It may sound counterintuitive, but keeping an extra defender at the third level is a better run defense against a team that counts on using either the read option or jet motion to put defenders in conflict.
One advantage of these two-high looks is how it leaves the defensive coordinator with several options for a defensive back to roll into the box. The example above shows the boundary cornerback blitzing the play-side of an Inside Zone, while the field defensive end is freed up to squeeze the cutback lane. This deception takes away the offense’s numerical advantage on the jet sweep and restores it to the defense.
In the front, Pry’s adjustment was assigning the defensive ends to “steal” gaps by playing two gaps instead of just one when they were the unblocked defender. The Rise technique was one way the Nittany Lions achieved this. The Athletic writer Ted Nguyen defines this technique: “When a defensive lineman rises, he stays square and squeezes the cutback lane to force a keep before exploding back outside to try and play the quarterback from inside out.”
Eventual 2020 2nd Round Draft Pick, Yetur Gross-Matos, earned 3.5 tackles for loss against Maryland as a gap stealer. Gross-Matos’ agility allows him to change direction easily, as he reads whether the play is a designed halfback run or a jet sweep. In the end, it didn’t matter which concept the Terrapins picked.
Another effective two-gap technique for defensive ends is the “Surf” (or “Feather”) technique which Nguyen defines: “When a defensive lineman feathers, he moves slowly toward the mesh point to force a ‘give’ read and then attempts to squeeze down to defend the cutback lane.” The tackle for loss Gross-Matos earned against a Power Read Shovel Option happens because he uses the Surf technique to be close enough to play either the run or the shovel pass.
An alternative approach to a complex versatile scheme is what Michigan State did against Canada’s offense: use the same front almost constantly. With the exception of their zone blitz package, the Spartans’ scheme almost always consisted of a 4-3 Over front and Quarters (Cover 4) coverage behind it. They aligned both safeties at such shallow depths behind the linebackers that they’re basically in the box. This wasn’t unique to matching up against Canada; this is their go-to against every team, Power I or Spread.
“Why Cover 4? We get nine men in the box,” current Pitt head coach and former Michigan State defensive coordinator Pat Narduzzi once said. “People talk about, ‘Man, we’re in an eight-man front.’ Well, we’re in a nine-man front.”
At eight to nine yard depths on most plays, deep safeties are legitimate threats to stop the run. The run fits of the Michigan State safeties on the perimeter are apt at taking away outside runs like a jet sweep or the play-call here: a Buck Sweep. In turn, the linebackers have the freedom to shoot gaps more aggressively when they identify the play is a run. Unlike Gross-Matos from before, the Michigan State play-side defensive end isn’t in conflict. His job in the example above is to blow up the blocking tight end, not disengage and passively play outside contain.
The fast-flow of all of these defenders together consistently held the Terrapins to just 29 total rushing yards and 0.9 yards per attempt. The Spartans thrived in pass defense too, allowing only 100 total passing yards. Quarters defenses in general match well against vertical passing plays, and this is why they’ve become so trendy in college football.
Michigan State’s version of Quarters is unique in how repetitive the coverage assignments are. The cornerbacks are generally in press MOD (man outside and deep) technique against the no. 1 receiver, the safeties are in a Robber assignment which require them to go man-to-man if the no. 2 receiver runs vertically eight-ish yards (otherwise, help the cornerback), and the outside linebacker has an underneath zone.
An incompletion on a Flood concept with Power play-action shows how the Spartans’ scheme is reliable against Spread run concepts and vertical passes alike. Solid press man coverage by the cornerback takes away the intermediate out route by the no. 1 receiver. The safety never has to play the flats in this coverage, so he can focus strictly on taking away deep routes like the go that the slot receiver runs. The pattern-matching—reading the receiver, as opposed to the quarterback—in Quarters coverage mitigates the effectiveness of play-action. You can’t bite the cheese on a run play if you see a receiver begin his route at the snap.
Pry’s emphasis on “stealing a gap” and Dantonio’s emphasis on nine men in the box were conservative compared to the tactics that Michigan defensive coordinator Don Brown used against Canada. While the college defense trend of the moment is lighter boxes and more Quarters coverage, Brown regularly sends the house on blitzes and leans on true man-to-man coverage. The distinct wrinkle in Brown’s specific game-plan for Canada was his use of odd fronts and interior stunts.
The Nut stunt (Nose loops into the A-gap in the opposite side of his initial alignment) is regularly used by defenses to counter zone-blocking schemes. This is because offensive linemen key on which defenders to block ‘pre-snap’ based on which gap they are in. If defenders exchange gaps, as the Nose and three-technique do here, blockers are vulnerable to missing their assignment. The novel wrinkle Brown adds is blitzing the linebacker who shadows the motion man into the B-gap, while the cornerback becomes the contain man. This additional inversion of assignments plays well against zone runs because it eliminates a cutback lane once the running back realizes the Nose clogged the play-side A-gap.
To set up another stunt, Brown uses an odd front with the middle linebacker standing at the line of scrimmage in the strong A-gap. This linebacker slants into the B-gap, setting a pick for Chase Winovich to loop into the A-gap. The center assumed he was “uncovered” (no down lineman in his gap), so he was fixated on getting to the B-gap to double-team Winovich. His over-aggressiveness is what creates a clear lane for Winovich to make a tackle for loss.
Like the Spartans, the Wolverines frequently kept their safeties shallow in a two-high shell; this is not something Brown regularly does though. When Canada called for a zone scheme paired with a jet sweep, the Wolverine safety to the side of the jet motion would roll down to that side and the other safety would play the zone scheme. This abundance of contain defenders frees up the defensive linemen to be so aggressive.
Nick Saban blended this tactic of rolling a safety into the box, with Pry’s tactic of using a two-gap defensive lineman to steal a gap. Alabama’s defenses regularly use a version of pattern-matching called ‘Rip-Liz’ that mirrors Cover 3 Buzz, except the cornerbacks have MEG or MOD assignments. They also have used two-gapping for many years, though less frequently.
For Canada specifically, Alabama usually rolled the safety to the side of the motion. From there, the safety and the play-side linebacker have fill and contain responsibilities, while the defensive end Anfernee Jennings (no. 33) two-gaps to free them up. Unlike Brown and Dantonio’s safeties, the Alabama safety opposite of the motion stayed deep. LSU’s passing game was more credible than Maryland’s, so that might explain this difference.
This safety rotation is useful against passing down-and-distances too. It attacked a vulnerability in Canada’s non-play action dropback pass protection schemes—they’re almost invariably half-slide protections. The blitz by Alabama’s nickel cornerback sets up the left safety to replace him in coverage. While running back Derrius Guice picks him up, that leaves a big gaping hole in the left B-gap. A Nut stunt by the Alabama defensive tackles leaves quarterback Danny Etling nowhere to run inside, while a speed-to-power rush by Jennings gives him nowhere to roll out.
The game that probably cost Canada his job at LSU though was the defeat to Troy. Troy uses the Tite front—a trendy method popularized by Iowa State in stopping Big 12 offenses. The Tite front uses just 3 defensive linemen who all are assigned an A or a B gap; the overhang defenders can stay outside the box and be put in less conflict. The down-side is less frequent or less conventional edge pressure on the quarterback, but the idea still worked well in this game.
When LSU attempts a jet sweep early in the game, we see how quickly the outside linebackers can flow to the play. Because the defensive linemen all clog the interior gaps, the Troy linebacker doesn’t have to go far to redirect away from the zone play-fake. This shortening of his path is highly effective in containing explosive runs.
Even when Canada turns to the Pistol, the pre-snap uncertainty of the zone run direction doesn’t inhibit Troy in this front. The two overhang defenders now set the edge instantly and prevent the offensive tackles from getting push downfield. Meanwhile, the MIKE linebacker is free to spill outside and, along with the box safety, clean up the zone run.
Even the most heralded defense that struggled against Canada’s offense, the Clemson Tigers, adopted the Tite front last year. In part, this played to their abundance of talented safeties and linebackers (especially Isaiah Simmons). But in any case, the combination of interior pressure and extra defenders in the seams is a logical strategy. It’s one that the next wave of offensive innovators will have to anticipate.
What will Canada’s influence on the Steelers look like?
This is still Randy Fichtner’s offense, unless there’s been a secret demotion we’re not aware of. I’m not trying to argue that the Steelers are suddenly going to be running Canada’s exact scheme. Canada was hired as the quarterbacks coach, and there’s no doubt that he was hired with the Steelers’ young quarterbacks in mind.
That said, I don’t believe that Canada is in Pittsburgh only to be a run-of-the-mill position coach. I find it hard to believe that it’s a coincidence 3 of the Steelers running backs—Conner, Samuels, and rookie Anthony McFarland—all played for Canada in college. Canada clearly made an impression on the Steelers several seasons ago when they began using his shovel option play-designs too. This is a coach who has the ear of his boss.
Mike Tomlin himself suggested this recently: “Matt’s gonna get an opportunity to work hand in hand with all the quarterbacks and help them in their growth and development and our readiness for the season. But not only that, man, he’s gonna bring some fresh ideas to us as a coach collective and probably bring some schematic ideas as well.”
I can foresee the Steelers’ offensive line being responsible for Canada’s greatest impact on the scheme. The Steelers have long been one of the most frequent users of gap-blocking run plays, namely Power O and Counter Trey. Right guard David DeCastro is one of the league’s best trappers on the edge. While Canada’s preference is generally for zone run schemes, his play-action pass protections that mimic gap run schemes are where I see potential.
This would require the quarterbacks to spend more time in the pocket, but the Steelers pass protection has been better than almost any at keeping them clean. As a unit, the Steelers ranked in the top eight in fewest QB sacks allowed in each of the last four seasons. Center Maurkice Pouncey didn’t allow a single one last year, while blindside left tackle Alejandro Villanueva only allowed two. Dialing up play-action concepts off gap run scheme looks would be a drastic shift in the Steelers’ passing game, but it fits very neatly into their existing run offense.
The additions of tight end Eric Ebron and fullback Derek Watt may also foreshadow this. The Steelers ranked 6th in 2019 (70%) and 11th in 2018 (69%) in 11 personnel frequency. It’s curious that salary cap space went to heavier personnel when they recently haven’t used them much. Meanwhile, another Hermsmeyer post shows that the most effective play-action passes are those that use less, not more receivers.
This is likely because most teams (including the Steelers) do not have a runner like Jackson at quarterback to make defenses still respect the run out of Shotgun or Pistol formations. Along with incumbent tight end Vance McDonald (a punishing blocker himself), Ebron and Watt may be spear-heading a shift to more 21 and 12 personnel this season, as in-line blockers and more. This seems like a logical compromise for the current roster construction.
My biggest concern comes from the Steelers backup quarterbacks, especially because Roethlisberger is not getting any younger and will be newly rehabilitated from elbow surgery. Mason Rudolph and Devlin Hodges had the two lowest air yards per completion in the entire NFL among qualified passers last season. Rudolph also had the widest spread between his air yards per attempt and air yards per completion, while Hodges was not far behind. That’s even more alarming because it suggests the Steelers backup quarterbacks weren’t reluctant throwers. They were just plain bad at throwing deep. Canada has used a lot of basic triangle reads common in West Coast offenses that Rudolph and Hodges could execute, but that isn’t what made his offenses special. The vertical play-action element did.
Roethlisberger still has enough arm strength to execute these play-action shot plays, but he reportedly favors playing out of Shotgun now. That isn’t necessarily a deal-breaker for Canada who used a healthy mix of Shotgun and Under Center formations. However, there is generally a drop-off in play-action in 11 personnel from Shotgun relative to Under Center.
Arizona Cardinals head coach Kliff Kingsbury believes that the difference in play-action efficiency between Shotgun and Under Center formations can be erased by a quarterback who’s a run threat. The Ravens provided some evidence of this last year with a 0.31 EPA/p in play-action passes from Pistol formations through mid-November, compared to NFL’s average play-action pass EPA in Shotgun (0.11 EPA/p) and Under Center (0.14 EPA/p). Roethlisberger is not a serious threat to run the ball and neither are his backups, so this is now a moot point for the Steelers. This is one reason why I thought they should have targeted Oklahoma quarterback Jalen Hurts with their second-round pick and planned for him to be Roethlisberger’s successor.
Yet, there might be no drop-off if the Steelers can stay in heavier personnel in Shotgun. According to Warren Sharp, play-action provides a significant boost with 12 or 21 personnel even in Shotgun. The Eagles, for example, rarely use Carson Wentz as a runner (despite Wentz’s history as a prolific runner at North Dakota State). In 2018, they still earned a 66% success rate and 9.8 YPA from Shotgun in 12 or 21 personnel. Across the NFL that season, play-action YPA improved from 6.5 in lighter personnel to 8.4 in 12 or 21.
This is why I believe Ebron is the key piece to making my suggested scheme changes work with Roethlisberger. The ex-Colts tight end lined up in the slot or out wide on 63% of snaps in his ultra-productive 2018 season (through November, according to PFF). His 40 receptions from the slot ranked third among tight ends that year, while his 7 touchdowns tied for second among all receivers in the slot. McDonald is a true in-line tight end, so he and Ebron should have clearly distinct roles if the Steelers opt for a large dose of 12 personnel from Shotgun.
Juju Smith-Schuster had a down season last year, but that is likely because of Roethlisberger’s absence. He was graded by PFF as the best receiver in the NFL on go routes in 2018. Roethlisberger counted on the go route even before Smith-Schuster’s ascendance; Antonio Brown ranked second that same season and excelled for years running it. That’s a route Smith-Schuster will have plenty of opportunities to run again if Flood becomes the staple I think it will.
Most of Smith-Schuster’s routes in 2018 came from the slot: 92 of 166 to be exact. It will be interesting to see where exactly Smith-Schuster and Ebron line up when they are on the field together, given that they both excel in the slot. The Trey formation where 3 receivers all are detached from the formation on one side, while an in-line tight end is isolated on the other, is the one that best fits what I think the Steelers are trying to do. It allows Roethlisberger to remain in Shotgun, gets McDonald on the field, and keeps Ebron and Smith-Schuster both aligned inside the numbers. The formation allows for three-level stretches galore in the wide side of the field, one-back power play-action passes with posts and crossers, and jet sweeps with McDonald as the arc blocker.
Another good option for the Steelers is what the Rams often do: align the receivers adjacent to the offensive line in bunch or stack sets. These condensed splits for the outside receivers retain the spacing of a slot alignment without actually being the slot receiver. If the Steelers decide against leaning on 12 or 21 personnel, this is how they can still use Canada’s ideas from 11 personnel. The condensed splits make the jet sweep more viable by packing the defense inward and leaving space on the edge. In 2018, the Rams ran the jet sweep more than anyone else in the league. The threat of the jet sweep helped them produce historic rushing efficiency on zone run plays (+0.10 EPA/p, higher than most team’s passing EPA/p). Am I the only one who thinks that sounds compatible with a Canada offense?
The biggest question to me for any of these approaches is the quality of the Steelers’ outside receivers. Canada’s offenses struggled most when the vertical passing game floundered and defenses could crowd the box. That’s also what happened to the Rams when the Patriots deployed a 6-1 front in Super Bowl 53 and other defenses copied that formula the following season. Smith-Schuster is a capable vertical threat from the slot, but this offense especially needs a vertical threat from out wide. Cornerbacks must play off the ball to let the perimeter run game, intermediate out routes, and shallow crossers thrive. Defenses must agonize on whether to keep a second safety deep.
The most likely candidate for jet sweeps is Diontae Johnson who admittedly isn’t a burner. Still, he broke a ridiculous 18 tackles after the catch last season as a rookie, third among all NFL receivers. Johnson should also receive a heavy dose of play-action targets because he ranked first among NFL receivers last season in average target separation, according to Next Gen Stats. The concern I have with Johnson is his ball security. His 6 drops and 5 fumbles last season raise questions about whether he has the physicality for the role I’ve suggested for him, or whether he’s a better fit for the slot too.
If James Washington plays up to his potential, he probably would have the biggest impact in this role. It’s true that Washington had success on deep routes late last season, even ranking fifth in the NFL on air yards per target (15.6). He also failed to catch 4 passes in any game in 2 years until the eighth game of his second season. The final 9 games last season were encouraging but his inconsistency is still worrisome.
This year’s second-round pick, Chase Claypool, is an athletic freak (6’4”, 238 lbs., 80-inch wingspan, 4.42 40 time) who will challenge Washington for reps. Claypool has promise as a red zone receiver. At Notre Dame, he caught 71% of his 46 targets in short areas (zero to nine yards). He tied for first among draft-eligible receivers with six touchdowns in that range. Personally, I think Claypool needs to improve his route-running between the 20s drastically before he seriously contributes. But if he does, I can see him becoming an isolation route weapon in 3×1 sets, paired with half-field reads like Flood on the trips side.
The Steelers are staking the potential of their revamped offensive personnel and scheme on a trio of unproven but intriguing receivers. If none of them pan out, get ready for lots of slow plodding drives. If one of them does, then the 12 personnel, play-action heavy strategy in Canada’s mold is their optimal strategy. If two (or all three) do, they can make any offense they want.
Liberty Gridiron 