The Pittsburgh Steelers are 2-6 in their past eight games.
Now, that isn’t fair at all. Before Pittsburgh was 2-6 in their past eight games, they were 11-0 in the preceding 11 games. Their defense was suffocating, their offense was efficient, and they were playoff bound. Seventeen-year vet Ben Roethlisberger captained that efficient offense despite limits on his arm strength: His average depth of target of 7.6 was nothing magical, but he was completing 67.5 percent of his passes and posting an expected points added of .212 per dropback (12th in the league).
Pittsburgh still made the postseason, but stumbled over the finish line with losses to Washington, Cincinnati, and Cleveland—and then Cleveland again in the wild-card round, in an astounding 48-37 loss. You remember: It was the game that started with a fumble-six and only got worse from there.
There were plenty of reasons for the prolonged stumble and thunderous face-plant that was the 2020 Steelers season, but there always has to be a fall guy, and the Steelers scapegoated offensive coordinator Randy Fichtner. A longtime staff member and close pal of starting QB Ben Roethlisberger, Fichtner couldn’t solve the many problems facing the Steelers offense—a deteriorating offensive line, a rotating stable of replacement-level running backs—and over time, the passing game crumbled alongside everything else. Roethlisberger’s depth of target dropped to 6.1; his completion percentage to 61.9 percent; his EPA per play to negative-.098, fifth worst in the league. The dwindling arm strength and shift at offensive coordinator foreshadowed a potential retirement for Roethlisberger, but he elected to return for another season, now under new offensive coordinator Matt Canada.
The book on Canada was that he could spice up the stale passing attack left by his predecessor. Fichtner had worked with Roethlisberger as his quarterbacks coach since 2010, and a large part of his appeal when he took over the coordinator job in 2018 was continuity and chemistry with Roethlisberger’s understanding of the offense. The offense was spread out and static, which let Roethlisberger dictate at the line of scrimmage and pick on any mismatches he favored. Canada represented a strong swing in the opposite direction.
Canada’s background in coordinating offenses comes from the college level. His rise began in Indiana, where he was coordinator of the offense from 2007 to 2010. Then, he was the offensive coordinator at Northern Illinois (2011), Wisconsin (2012), NC State (2013-15), Pittsburgh (2016), LSU (2017), and Maryland (2018). That transience has made Canada an interesting figure in football circles, but his offense isn’t a perfect fit in Pittsburgh. Canada wants to move people before the snap, he wants to put his quarterback under center, and he wants to run play-action. Historically, Roethlisberger wants to do none of that. The challenge for Canada and Roethlisberger entering the season was to find a happy balance.
Whatever they’ve found isn’t exactly working yet. Roethlisberger’s EPA per play (.071) and completion percentage (62.5 percent) are both closer to the late-season stumble than the early-season surge, even as his depth of target has jumped to 8.2.
On the surface, Canada is pressing all of the right buttons, like pre-snap motion, for example. Pre-snap motion is a lauded feature of many prominent offenses, and increasing motion rates is beneficial for offenses in limbo, like the Buccaneers’ offense in 2020. If there’s a player who can move before the snap, Canada will find that player, and trust me—he’ll move him.
This is just Matt Canada flipping the entire O-Line/TE, putting the RB in the wing, then running him on a jet motion. Fun. pic.twitter.com/Cf2mba5jV3— Ollie Connolly (@OllieConnolly) October 23, 2017
Pre-snap motion is broadly helpful to NFL offenses, but individual players and play-callers can struggle with it. Even after the success Tampa Bay had moving players pre-snap, offensive coordinator Byron Leftwich denied the idea that his group would continue working with new pre-snap shifts, saying “the offense is based on the quarterback.” While there is still plenty of motion in the Bucs offense, motion isn’t a panacea; it isn’t a no-loss proposition. Leftwich is coaching under Bruce Arians, who recalls that Colts QB Peyton Manning never liked players in motion during his time in Indianapolis. “Peyton Manning never wanted anybody in motion, so each quarterback is so different [with] what they want the motion for.”
Roethlisberger clearly falls into the camp of quarterbacks who are uncomfortable with pre-snap motion. In 2020, the Steelers moved a man on 42 percent of their snaps (20th in the league). Roethlisberger spoke during the season about the use of motion and play-action as something he was still trying to acclimate to; Michael Lombardi of The Athletic said that Ben would never like play-action drops from under center, another common characteristic of the Canada offense, because they forced him to turn his back to the defense. Fichtner said something similar during the season, when he said “we evaluate what can be done from a play-action standpoint from a protection standpoint of Ben Roethlisberger first. He is going to always be most comfortable in dropback pass. He can see in front of him. He can see his sides. He can be prepared for sight adjusts and hots and things like that.” Years of data shows that Pittsburgh doesn’t use play-action often and doesn’t find much efficiency when it does.
The Steelers are in their own tier when it comes to play-action usage and success over the last five years.— Ben Linsey (@PFF_Linsey) May 28, 2021
Since 2016 (@PFF):
12.3% play-action rate (last)
6.1 yards per play-action dropback (last) pic.twitter.com/iowi2VD2D1
While Fichtner is speaking specifically about play-action, pre-snap motion can have the same deleterious effects for a cerebral pre-snap passer. When you move an offensive player, the defense moves as well. That can give you some information about the defense’s plan for that snap, but it can also give the defense an opportunity to realign and disguise as well.
Early returns show that Roethlisberger has been willing to evolve in his 18th season as a pro. The Steelers have run a play-action fake on 25 percent of their dropbacks this season, per PFF (up from 10 percent in 2020). The Steelers have also run more plays from under center, though to a lesser degree—Roethlisberger has only five passing attempts from under center this season, all of which came with play-action. That puts him at 25th in the league; he was 38th in such dropbacks in 2020. And, somewhat strangely, on pre-snap motion, the Steelers have actually dropped from 42 percent of snaps in 2020 to 35 percent in 2021.
It’s still early in the Canada era, so we should pause before we boldly proclaim that this is the final and ideal iteration of the Canada offense with Roethlisberger at the helm. There’s more time to find an optimal rhythm in play-calling and play designing. With those qualifications and excuses put aside, it’s clear that play-action and pre-snap motion haven’t and won’t make Roethlisberger’s arm magically stronger.
Let’s take those increased play-action reps: Roethlisberger is currently averaging 6.7 yards per attempt on play-action, which is the same as his straight dropback reps. But he pushes the ball downfield less on play-action, with an average depth of target at 6.1 as opposed to 8.9 on straight dropbacks. That’s the ninth-shortest depth of target on play-action in the league.
Remember, play-action fakes take time, and the time they take is what allows receivers to get downfield and for passers like Ryan Tannehill, Josh Allen, and Tom Brady to attack those windows. But Roethlisberger’s limited arm strength prevents the Steelers from drawing up these ideas.
Here’s a common downfield passing concept in the NFL: 989. It’s Air Coryell terminology: The 9 stands for the downfield go route, while the 8 stands for the post route from the slot receiver. Remember the numbered route tree you learned in peewee football? Same basic idea here.
This is a three-step drop from the gun—Roethlisberger’s preferred way of playing ball. He initially wants the in-breaking route in the middle of the field, but does not feel confident that he can lace the ball into that shrinking window, especially with the safety lurking. So he looks to the far sideline and loads up for a downfield throw—the correct read, as he has a one-on-one opportunity—but then elects to holster the ball to try to break the pocket. He’s sacked.
They went for max protect 989 and it went *exactly* as you'd expect pic.twitter.com/HGnv7Q5JxK— Benjamin Solak (@BenjaminSolak) September 21, 2021
The problem isn’t that there was no play-action fake attached. The problem isn’t that there was no pre-snap motion to move the defense around. The problem is that Roethlisberger had a one-on-one go route, 25 yards downfield, on the far sideline, and he knew he couldn’t make the throw.
Now, don’t get it twisted. Roethlisberger may have made the throw with a clean hitch into the pass and a giant Canadian waiting at the other end: Chase Claypool’s large catch radius has been the saving grace of Pittsburgh’s downfield shots for almost a calendar year now. He hit Claypool on a deep pattern against the Raiders that required a good deal of velocity and distance. Arm strength isn’t a certain proposition, in the same way that you’re better at the gym on some days than you are on others. Aside from miracles of nature like Brady and football-throwing cyborgs like Allen, arm strength can vary based on platform, release mechanics, pressure, arm angle, and just general wear and tear. Roethlisberger’s diminished arm strength limits the Steelers’ offense as a whole, even if there are still flashy moments on individual plays.
Folding play-action and pre-snap motion into the mix simply doesn’t solve this problem. Again, these concepts have been universally helpful throughout the league, but there are some things that cannot be helped. Eating better will help you live both a longer, healthier life—but it won’t keep you alive forever. With an under-center dropback, motion at the snap, and a play-action fake, Roethlisberger still doesn’t have the velocity necessary to drive this ball into an extremely tight window.
This isn’t a throw that many quarterbacks make. Las Vegas cornerback Casey Hayward plays it well, and even with a faster ball, poor placement makes this a much easier interception. But you’d like for your quarterback to be able to throw this wheel route in two different ways: with touch if it’s man coverage and with velocity if it’s zone coverage. Roethlisberger has only one of those tools in his belt.
Canada can try to find easier wins with pre-snap motion and play-action. The Steelers’ screen game is humming right now (10-for-10 for 72 yards), and likely deserves more attention. But it’s not like the Steelers don’t know what concepts to run for Roethlisberger—they’ve been running all the same quick-game ideas since last year, and we saw the diminishing returns on that offense as defenses settled into press coverage or flooded the short area of the field with zone defenders. The windows got smaller, the slants and drags were contested, and the passing game fell to shambles.
It’s good that Roethlisberger and Canada are working to find a happy medium. I’m just not sold that the happy medium will create a higher ceiling for the Steelers’ passing game. Nothing about the Steelers’ passing attack will scare defenses, and no number of players they move before the snap or fakes they execute to the running back will change that. In the NFL, you either have a quarterback and you’re a competitive team, or you don’t and you aren’t. We saw what the Steelers looked like down the stretch last season against competitive teams. Why would we have any reason to expect any differently from them this season?