Super Bowl Recap: Defense
![Lot of this. Coulda used more. [Scott Strazzante]](https://images.squarespace-cdn.com/content/v1/5e40f3a094d5fa3ebbbedb09/1581316192989-25ASIN6EGI68IPNF7ZI9/2880x0.jpg)
Lot of this. Coulda used more. [Scott Strazzante]
This is your regularly scheduled reminder not to trust the media.
As predicted, a series of hot takes have been flooding in since the Super Bowl ended, and -- given the end result -- their headlines and narratives have been just as predictable and off-base as you could expect. In general, football is hard to tackle for your standard sports journalist, largely because said journalists crave a five word description that can sum up the entirety of a 60 minute game of 11-on-11 that requires the active regular involvement of 30+ players per team, a dozen coaches per side, and a never-ending series of nuanced complex schematic decisions. The local writers are the ones who typically know their shit the most cause they're in it all year and are more team and sport specific. But on a national scale? You've got some Bill Barnwells and people on The Athletic -- sports people who actually know what they're talking about -- but in general, football journalism has become a lot like every other form of journalism. Articles built backwards from their trend-worthy titles. Click-bait prorated into a few hundred words. Empty calories.
And if John F. Bachelors Degree in English with a minor in French postmodern literature from Emerson College (Class of 2011) is fine with setting that bar, what do you think @FortWhiners69 and @RIPNateDogg510 have to say about it? Please, by all means, let's turn to @KansasCityQueefs and @TylerThigpenEra2008, because they really have some outstanding football analysis when they're not trying to sell me dope discount Ray-Bans at one tenth market price as long as I buy now.
All this to say, everyone has their opinions. But when those opinions are simply misstatements of the facts — which in itself is a much larger problem that is much more damaging to realms much more important than football — they probably qualify a bit more as bullshit than they do as opinions. So don’t feel bad if you shut them down. Because people being wrong who are adamant that they're right is the backbone of the most dangerous of breed: the unreasonably confident incompetents. And those are the guys who really mess things up for everyone. Or, you know, go on to ref the Super Bowl after blowing the Saints-Rams game last year.
I'm not salty. You're salty.
Alright, let's get to the recap. Defense first.
DEFENSE
According to script… until it wasn’t. As expected, we ran more two-high looks, gave the Chiefs lighter boxes to dare them to run, mixed up our coverages, and leaned on our overall team speed and disruptive DL. For the first three and a half quarters, it worked. Then, it didn’t. We’ll try and unpack why later.
Zebras (usually) have excellent eyesight. I already wrote up plenty about the officiating in this game. It’s almost impossible not to mention it when discussing some of the crucial plays in this game because so many of those plays were directly affected by the refs. But I will try to keep that discussion to a minimum.
Dudes playing like dudes. Our top dudes along the defensive line continued their dominant stretch, controlling the line of scrimmage and pressuring Mahomes into an inconsistent performance and some uncharacteristic mistakes. But none were more disruptive than Nick Bosa, our would-have-been MVP had we won this game.
Bosa dominated from the jump, punking Eric Fisher on the very first pass play of the game and tallying 12 pressures and 10 QB hurries in the contest. He was an unstoppable force for the first three quarters of the game and — while our pressure as a whole slowed down some in the fourth — could have (likely would have) put the cherry on the top of an MVP performance had he not been blatantly held after beating his man on the third-and-15 conversion to Tyreek Hill.
Bosa was so dominant that his impact on the Chiefs’ play-calling became obvious in a hurry. Case in point: Mahomes’ first pick of the game.
Not only did they roll the pocket away from Bosa, splitting their field in half, but they still dedicated THREE GUYS to blocking him.
Dude is an absolute force.
Other guys played like… well… While on paper this really didn’t seem like the game that we’d miss our pluggers in the middle (DJ Jones and Jullian Taylor), our lack of girth was apparent against the run. Against the Vikings and Packers we could get away with focusing on the run, knowing that they needed their run games to slow down our DL and set up their passing attacks. But against the Chiefs, we (reasonably) leaned more on two-high sets to stop the pass, which put more pressure on our front six to execute and win their one-on-one matchups.
For the most part our edge players delivered, but our non-DeFo tackles proved a weak spot en route to an 18 carry 104-yard rushing performance by the Chiefs’ tailbacks. Our subpar play on the interior was evident in PFF’s grades for the game (Day - 46.6, Thomas - 33.8, Mitchell - 29.3). There were just a few too many plays like the one below (featuring Thomas and Day):
I know they were dealing with double teams, but you just can’t allow that kind of push and expect to have a strong showing against the run. And while the Chiefs’ yards-per-carry was only an “okay” 3.8 when you take out the long run after the game was over, our play on the interior allowed the Chiefs to chip away for a lot of small consistent gains on the interior. And in a game as close as this one, those little gains can add up.
Option game. Reid dialed up a number of different ways to attack our edges with the speed option game (like on the game’s first play from scrimmage), namely by pressuring the solo side of trips formations to ensure he could outflank us. It wasn’t a gamebreaking development, but it worked enough times on crucial downs to be worth mentioning. Reid knew he had to keep some semblance of balance to attack our defense, so he did so by stealing a gap with the option game.
Due to this and the above mentioned issues on the interior, the Chiefs did just enough to get some first downs on the ground and keep us honest. Which was likely their entire goal of the running game in this game.
Having boundary issues. While Sherman and Moseley played outstanding this year, both of them struggled in this one. Moseley was at least partly complicit in two of the Chiefs’ big completions over 20 yards and Sherman really struggled in man coverage, allowing more yards in this game than he had in any other game this season.
The speed of the Chiefs receivers really got to him, including on the deep shot to Sammy Watkins where he got beat on an outside-in fake at the LOS and on the long third down on the Chiefs’ very first scoring drive, where Sherman’s slip on the press of Mecole Hardman created a domino effect that pulled Tartt off his robber/spy duties to pick up the rookie, and which — in part — allowed Mahomes the open lane to scramble for a fourth-and-one. The next play they would convert with that old Michigan play that people won’t stop talking about. They’d score a touchdown three plays later.
Kwon. I’m not gonna be too hard on a guy who came back crazy early from a torn pec because he wanted to help his teammates in the playoffs, but he did not play well at all. He had a handful of bad missed tackles, wasn’t particularly sharp or confident in coverage while giving up a handful of receptions, and stood out for all the wrong reasons. This despite playing less than a quarter of the defensive snaps. According to PFF, his 24.6 grade was the worst of any player on the Niners roster.
Masters of Disguise. We did a great job of disguising our coverages for the majority of the game. messing with Mahomes’ pre-snap reads just enough so that our pass rush could flush him into a bad or ill-timed throw. Take this first down in the third quarter, deep in our territory:
We’re in a single-high look with Moseley squared up on Sammy Watkins at the top of the screen. So it looks like (at least on Moseley’s side) some kind of man coverage.
After sending Tyreek Hill in orbit motion — both as a coverage indicator and as a means to work their four man flood concept — the Chiefs see K’Waun Williams follow him across formation, doubling down on the idea that we’re in man.
On the snap, Mahomes is expecting he’ll get a natural rub from his two vertical routes coming out of the backfield, opening up Watkins on the slant. But despite our nickelback following Hill across the formation, the Niners aren’t actually in man. They’re in a three deep zone.
By the time Mahomes has completed his play fake and looked up, our boundary corners have taken the deep routes, Warner has moved off his low wall and is now double teaming the seam route with K’Waun, and Greenlaw is working to high wall Watkins, who has adjusted his slant to a deep settle route upon reading zone coverage.
Meanwhile, with Kelce going across formation behind the LOS, Tartt can play a full “Robber” role, looking to undercut anything underneath and inside and keeping his eyes on Mahomes as a spy in case he scrambles.
The result? A confused pre-snap read, no one open, and a DeFo sack.
Well, after he was blatantly held that is.
Because last time I checked “wrapping your hands around a defender’s waist while you grab them from behind like you’re taking photos at prom” is not a legal blocking technique. Except for, apparently, in this game.
Alright, I’ll stop.
…
No I won’t. Call a fucking hold.
War of attrition. But… there’s only so many sound coverage combinations you can show someone, and when you give a talented passer and play caller too many shots, they’re likely gonna start figuring it out and hit some. Particularly when someone makes a mistake. Like on the third-and-15 conversion in the fourth quarter that everyone and their mom is having a boner over (yes that’s right, their moms have boners).
Here’s how that play unfolded, courtesy of Ted Nguyen at The Athletic.
In the Super Bowl, the 49ers played Cover 3 Buzz. In Cover 3 Buzz, the linebacker will run with the No. 2 receiver vertical for a distance, but it is the corner’s or free safety’s responsibility to take him deep.
Hill lined up as the No. 2 receiver and tight end Travis Kelce lined up as the No. 3 receiver. Kelce ran a deep crosser and Hill ran right at Ward on what looked like a seam. The play was designed to look like verticals, initially.
However, instead of a fade, Watkins ran a dig on the outside. Cornerback Emmanuel Moseley, who was supposed to stay in his deep third, made the fatal mistake of following Watkins. Ward’s job is to defend the middle of the field, so when Hill broke off his seam to the corner, he had no shot of making the play.
Top-tier quarterback + top-tier play caller is always gonna get theirs at some point when they can protect it. Add in a mistake by a cornerback trying to drive on a dig at the sticks and a pass rush that didn’t hit home in part because of a blatant missed hold and that’s how you start giving up big plays in the passing game.
Just how it is sometimes. While not a revelatory answer nor a satisfying one, it seems fitting in this case. There’s no single reason why our defense collapsed with half a quarter left in the Super Bowl, and when you look at their performance objectively it’s not really as much of a collapse as it may seem.
Let’s eliminate the long TD drive when the game was — for functional purposes — already over and look at their two fourth quarter scoring drives to take the lead:
Drive #1 was 12 plays for 83 yards and took 2:34 seconds. But 65 of those 83 yards happened on two plays, both on 3rd-and-10+: the 44-yard completion to Hill which should have been a hold on Bosa and the 21-yard pass interference call on Tarvarious Moore that put the ball on the one-yard line. So otherwise you’re looking at 10 plays for 18 yards…
Drive #2 was 7 plays for 65 yards in 2:20. An early first down came when Bosa narrowly missed Mahomes and blocked Dee Ford (who had also beat his man) in the process. Then 38 of those yards came on Sherman getting beat deep in man coverage by Watkins. To cap it off, the five-yard score was on a blatant offensive pass interference on Travis Kelce (on once again, a third down).
Football is a sport where a single missed assignment or call can start a chain reaction that leads to an entirely different outcome. It’s part of the randomness that comes from the sequential nature of the game and the far fewer possessions that football has versus a sport like basketball or hockey. Sure, in retrospect you can say the Niners shouldn’t have been in man coverage on that big play, but putting your All-Pro cornerback on their third option when you have trips on the opposite side is not a crazy or egregious play call.
A lot of shit had to go wrong to lose this lead, and the bad breaks and bad calls just seemed to converge in this one.
So what’s next? While the Niners, as much as any squad, have the right blend of coaching, management, locker room leadership, and young talent to theoretically become a consistent contender, a lost Super Bowl opportunity is always one thing. And that’s a kick to the balls. Potentially even more so for teams that rely on a dominant defense.
Plus further evidence from Josh Hermsmeyer at 538:
Defensive performance is just much harder to predict and repeat than offensive performance. The rules (in general) are built to help offenses, it’s harder to hide a weak link on defense, and the big negative plays and takeaways that are so important towards defensive performance are — as shown in the chart above — some of the hardest plays to predict on a year-to-year basis. That’s especially the case for takeaways.
In 2018, we ranked 25th in Defensive DVOA, in part because our historically bad takeaway numbers had us last in the league at generating turnovers. In 2019, we jumped to 5th in Defensive DVOA and 6th in takeaways. While our off-season influx of talent may make that jump seem predetermined, the ebb and flow of team takeaways is much more random than that.
Take a look at this graph (which I totally just made to test out the graph building functionality of Squarespace). It shows the six teams who either led the league (1st) or were last in the league (32nd) at takeaways during the past three years.
Takeaway Rankings by Year (2017-2019)
Every single one of those teams ranked 22nd or worse AND 6th or better in takeaways at some point… in only a three year span. And of those six teams, the Ravens were the only one to get into the top ten more than once.
Even if our defense continues to be strong, and there is every reason to believe it will be, it’s worth expecting our takeaway numbers to move closer to the mean. Which would — in theory — move our overall defensive output closer to the mean as well.
But if there’s hope for continued defensive dominance, it’s likely in our pass rush. This year we led the league in hurry rate and finished second in sack rate (both important, influential, and potentially repeatable statistics for offensive disruption). But QB hits — a statistical category where we’re oddly middle-of-the-pack — has proven to be one of the most consistently predictable performance stats (defensive or otherwise) on a year-to-year basis.
Once again, from Josh Hermsmeyer at 538:
Still there is some hope for lovers of the three-and-out. While rare, there are plays a defense makes that do tend to carry over from year to year. One of the most stable defensive stats is hits on the quarterback, which has a relatively impressive year-to-year r-squared of 0.21 — better even than total offensive DVOA, which is the gold standard for stability in team metrics. Quarterback hits include sacks — 43.5 percent of QB hits end in a sack, and those by themselves tend to not be predictive — but also plays in which the passer is contacted after the pass is thrown, and that contact is incredibly disruptive to a passing offense.
I’m picking and choosing for this graph, so take it with a grain of salt, but I tried to select teams who placed in the top ten in QB hits during the 2017-2019 stretch and had the same scheme and/or defensive mind in place for all three years. Then I added the Niners for context. And the Raiders (lol).
The results are encouraging.
QB Hit Rankings by Year (2017-2019)
That’s four strong defensive teams, who — despite having differing schemes and, in some cases, a rotating door of talent — have been incredibly consistent in terms of generating QB hits.
And while our QB hits have stayed near league average the past three years, I would guess in our case that that’s an anomaly. Like many other statistics, our QB hits took a dip in the back half of our regular season before picking up again once our defense got healthy (we registered 8 or more QB hits in three of our last four games, including the Super Bowl). And while each year has teams and players whose pressure is overstated because they convert hits into sacks at an abnormally high rate, leading (or near-leading) the league in both hurry rate and sack rate would imply that the QB hits will soon be on their way.
So perhaps it isn’t all doom and gloom. We’re young, well-coached, and — if we can keep up our pass rush — have the means to suit up another dominant defense in 2020. But it’s unlikely we’ll have the turnover luck we had in 2019, and in order to offset that and build a perennial contender, our offense will need to improve.
That’s next time.