I like mysteries. Glass Onion was released to Netflix over the holidays, and while I didn’t like it quite as well as Knives Out, I enjoyed the conceit of a mystery with layers that don’t really matter because you can already see right to the middle, like the titular glass onion. The Vikings last week were a glass onion. No one outside of Minnesota was surprised they lost or that they lost by quite a lot. The Vikings’ shortcomings have been on full display all season via DVOA, EPA/Play, point differential, and gambling lines. Much like Miles Bron, the main character of Glass Onion portrayed by Ed Norton, the mystery isn’t so much about the downfall as it is how they got those 12 wins in the first place.
This week, we have the Detroit Lions. I spend every week looking for little statistical patterns of every Green Bay Packers opponent for strengths and weaknesses to figure out at least a few examples of what may happen in the upcoming game. While the Vikings are the Glass Onion of the NFC North, the Lions are much more Murder on the Orient Express, where the facts are not just confusing, but self-contradictory; where everyone did it and no one did it, and every motive comes with an annoyingly airtight alibi.
Let’s begin with Football Outsiders metric I use the least: Variance. Variance simply measures how consistent a team is from week to week. The most consistent team this year has been the Washington Commanders, which is actually pretty amazing given their quarterback nonsense. The Packers were right about in the middle at 18th. And dead last, at 32nd, we have the Lions, who were statistically speaking, crazy pants on a week to week basis.
What exactly does this kind of variance look like? One example is their matchups against the Bears and Vikings. The Bears on the season had a terrible total DVOA of -26.5%. The Vikings were not much better at -14.2%, and indeed, they were the two worst teams Detroit played this year in terms of DVOA. In Detroit’s Week 3 loss to the Vikings in Minnesota, they posted a rather atrocious -22.1% DVOA. In their week 14 victory over Minnesota in Detroit, they reversed the trend to post a 16.4% DVOA. Detroit did manage to sweep the Bears, but their Week 10 31-30 win in Chicago wasn’t a thing of beauty, as Detroit’s DVOA for the game was -20%. Last week in Ford Field, they destroyed Chicago, putting up their 2nd best performance of the season at 70.2%. (Against Jacksonville they posted a DVOA of 80.4%).
That’s a lot of variance! Now, you may be thinking, Paul, is this just a home-road thing? And yes, there’s a little bit of that for sure, but it’s not as big of an effect as you might expect. Detroit is better at home as most teams are, finishing 5th in home DVOA with an 18.9% average, but they’re hardly atrocious on the road, finishing 12th overall with a 3.6% DVOA. (Note: Most teams are worse on the road, and so average is lower. There were 10 teams this year that were better on the road. Baltimore, for some reason, was second on the road with a DVOA of 16.1%, but crashed to 20th at home, with -.9%.) The Packers have almost the same home/road split as the Lions, ranking 8th at Lambeau (14.1%) and falling to 13th on the road (0.6%).
So yes, Detroit is worse on the road, but not weirdly so, and while they did get destroyed at New England (-47.3% DVOA) and Dallas (a season-worst -60.8%), they played very well at the Jets (24%) and the Giants (39.3%). Maybe they just like New Jersey. And while Ford Field was generally friendly, the Lions were crushed by a middling Seattle team (-47.3% DVOA, Detroit’s 2nd worst performance of the year), and had a lackluster performance against a pretty meh Miami team (-6%).
I hate variance. Variance is just a statistical word for nonsense, and sometimes when there’s variance, there is very real nonsense at its heart. One of the reasons I was so surprised to see the Commanders so low in variance is that their coach nonsensically played their worst quarterback in an elimination game for no reason. Such ridiculous behavior will throw a wrench into any refined system, but with the Commanders, I think what we really see is that ridiculousness permeating the entire season in less obvious ways, which is why they were so consistent. And while parsing Detroit is tricky, there are a few patterns hiding in that variance, which we can find by interrogating the data further.
The Detroit “randomness” has three key components at its heart. There are a few very different ways to beat the Lions, and every one of those ways is very different from the others. Some are skill based, some are luck based, and few are in Detroit’s direct control. For example:
1. Opposing Pass Defense
The Lions have a good offense, and I think it’s important to establish that upfront. They scored over 30 points seven times, while Green Bay managed that feat only three times (all three coming after their week 9 loss at Ford Field). Getting into a shootout with Detroit (6th in offensive DVOA on the year) is a dicey proposition, but they weren’t immune from a talented opposition. The Lions faced six opponents with an above average pass defense on the season: Philly, New England, Dallas, Buffalo, the Jets, and the Packers. The Pats, with their -13.5% pass defense (remember that negative numbers are better on defense), actually shut out the Lions, while the Jets held Detroit to a single offensive touchdown. The Lions scored only 15 on the Packers despite plenty of help from the Lions defense via turnovers. The Cowboys dominated the proceedings and held them to just 6 points. Buffalo held the Lions to 25 points on Thanksgiving while also getting to Goff for a safety.
Finally, we have the Eagles, who, despite having the league’s best pass defense on the year, gave up 35 to the Lions. What happened? Weirdness, of course.
The Lions opened with a nice drive highlighted by a 50-yard D’Andre Swift run. While the Eagles were first against the pass, they were 20th against the run, and it was a major problem for them all season. Jamaal Williams would punch in a one yard touchdown.
The next Lions drive was a 3-and-out. The one after that was a 3-and out. The one after that was a 3-and out. And the one after that ended on a Goff Pick-6 by James Bradberry. The Eagle pass defense was as dominant as ever during this stretch, giving Philly a 21-7 lead. Philly would go into halftime up 24-14 and scored on their first two drives of the second half. With 1:41 remaining in the 3rd quarter, they were up 38-21, and more importantly, their lead was never seriously in jeopardy. Detroit would cut the lead to 3 with 3:51 remaining (helped by an 11 yard DPI penalty against CJ Gardner-Johnson on 3rd and 10), but the Eagles just ran out the clock. Philly had a terrible time stopping Williams and Swift on the ground, but mostly, they didn’t care. They got out to a huge lead and as teams do, they got conservative. When they cared to play pass defense, Detroit was completely incapable of passing. You can see it in Goff’s completion percentage, as he was just 21/37 for a mere 215 yards. Swift and Williams, on the other hand, combined for 172 rushing yards on just 26 carries.
If you can shut down the Lions’ passing attack, you give yourself a great chance to win the game. The Lions went 2-4 in games against above average pass defenses. The Packers and Jets were their two victims, and without a Kalif Raymond punt return touchdown against New York, they lose that game as well.
2. 4th Down Aggressiveness
Early in the season, Detroit lost a heartbreaker to Seattle, 48-45. Seattle combined an efficient passing attack with capitalizing on Detroit’s poor rushing defense with a 33-carry, 235 yard, 3 TD performance on the ground. Capping it off, on the first play of the second half, Seattle’s Tariq Woolen pick-sixed Goff for a 40-yard score. It was an exciting game that wouldn’t have been close at all without three 4th down conversions (on three attempts) from Detroit, which allowed them to hang around. The first, a fake punt pass from Jack Fox to Quintez Cephus, would lead to a Detroit field goal. The second, a gutsy call on 4th and goal from the Seattle one yard line with just three seconds left in the half, resulted in a Jamaal Williams touchdown run. The final conversion, a 17-yard pass to Tom Kennedy during a furious comeback attempt, led to an eventual 4-yard touchdown pass to T.J. Hockenson.
The Lions are ridiculously aggressive on 4th downs, having gone for it 35 times on the season, behind only Cleveland (40) and Arizona (39). They’ve converted 18 of them (51.4%) of them, and those conversions often have a huge impact on their points scored in a given game. The Packers also go for it routinely, having done so 29 times, but they’re not nearly so adept at converting, succeeding just 11 times (37.9%).
That aggressiveness is well-founded, and analytics people spend much of their Sundays urging teams to go for it on 4th and 1 or on their opponent’s side of the field. We know that in the aggregate, over time, going for it is almost always better than a punt from inside the 50 or a short field goal. But 4th down plays are also extremely high leverage plays, and the success of failure can swing an individual game, leading to some extra variance. The Lions went for it 6 times against the Patriots and converted 0. It was a huge factor in that game, and one of those failed attempts turned into a sack-strip-touchdown of 59 yards by New England’s Kyle Duggar. The Lions were ultimately shut out, but they were close to scoring on several occasions. In just the fourth quarter they failed on attempts from the New England 5, 31, and 18 yard lines.
The Lions actually went 0-2 on 4th down in the Packer game, failing on a pass to Tom Kennedy at the Green Bay 7 on their first drive of the game and again in the 4th quarter from the Green Bay 43. Unfortunately, the Packers had 4th down problems of their own, going just 1-4 and failing to turn either Detroit failure into points.
4th down aggressiveness is a good idea, but it can inject variance into the game, especially when your conversion rate is extreme in any one game, and so predicting a Lions performance means predicting just how they will fare when a 4th and short inevitably comes up. Lastly, I should point out that Detroit is adept at running trick plays, and it’s a part of their arsenal to beware of.
Let’s go back to the first meeting between the Lions and Packers in Week 9 once again, with four specific plays in mind, and let’s start with Detroit safety Kerby Joseph. Per PFF, Joseph ranks 61st out of 88 safeties. He’s not exactly a shining star among defensive backs, but in this game, he put up quite the stat line. In the first quarter, the Packers had the ball on the Lions’ five yard line after an impressive 6-minute drive. The Packers went to one of their patented garbage red zone RPOs, and Rodgers’ pass, intended for Allen Lazard on a slant, was batted into the air. Joseph just happened to come down with the ball. While Rodgers has seen an alarming uptick in batted balls this year, most batted balls end up on the ground. It was first down, and that pick saved at least 3 points.
Joseph would get Aaron again in the 3rd quarter just two plays after Jaire Alexander intercepted Jared Goff, essentially erasing the biggest play of the game for the Packers. Rodgers attempted to force the ball deep over the middle to Robert Tonyan, and Joseph made a nice play closing on and catching the ball. It was a questionable throw as Tonyan was well covered, but I think I know why Rodgers threw it, which brings us to our third play, which occurred one drive earlier.
On that drive, the Packers were in business after an 18-yard completion to Christian Watson brought them to the Detroit 43. On the next play, Aaron Jones was stuffed for a 6-yard loss, and while the Packers would get 5 of those back on second down, they faced a 3rd and 11. On that play, Samori Touré broke wide open over the middle for what should have been an easy score. Unfortunately, Aaron undershot him by the smallest of margins, and Kerby Joseph, once again, made a nice athletic play to bat the ball away. It really should have been a touchdown, and I think on the Tonyan pick they tried a similar concept with Tonyan in place of Touré, but the Lions were ready for it.
I think at least a few of these can be blamed squarely on Rodgers’ thumb. The underthrows and overconfidence in his arm defined the broken thumb era of this season, and if healthy, I’d wager the Packers have another touchdown or two in this game, which brings us to our 4th play, a 4th-and-1 pass to kick off the second quarter. The Packers decided to run a trick play in David Bakhtiari’s first game back, targeting the left tackle with a pass. Bakhtiari was even open, but Rodgers made a poor throw that was picked off by Aidan Hutchinson.
All of these were fairly bizarre plays, unlikely to repeat themselves. The Packers would fail on 4th down several times in this game as well, including on a video overturn of an Allen Lazard catch. They outgained Detroit 389 to 254 yards, but lost the turnover battle via typical turnovers and failed 4th downs, and that was that.
The Lions held the Packers to 9 points, but I think it’s important to highlight just how bad they are at preventing yardage, and how large a factor turnovers are in their success. Saying that the game comes down to turnovers is among the most clichéd aspects of football, so let’s do a bit better. Jared Goff has thrown 7 interceptions on the season, and none since the first Packer game. Except for that Packer game, the Lions have lost every single game in which Goff has thrown a pick. Of course, it helps when Rodgers throws three picks on the other side.
In addition, the Lions have been outstanding at not turning the ball over, and have played completely clean games eight times this year. Their opponents, on the other hand, have only done so twice. A good chunk of their total season turnovers came against Dallas where they had 5. Here’s the thing: Jared Goff, while not a bad quarterback at all, usually isn’t this careful with the ball. This season, only Jalen Hurts and Daniel Jones (!) have a lower INT%. Goff was, to his credit, also careful with the ball last season, but even then, his INT% was still higher than it is now, and he was REALLY checking it down last year (9.8 Y/C versus 11.7 this year). Goff should be praised for the brilliant season he’s having, but when you see numbers like that, there’s usually some luck involved, and indeed, opponents have dropped their share of interceptable balls.
Updated luck dashboard weeks 1-17.— Tom Bliss (@DataWithBliss) January 4, 2023
Largest Net Win Probability Added play in Wk 17: SF 41-yd missed FG that would've won game in regulation vs LV. pic.twitter.com/5MefUrUNqn
Turnovers are a mishmash of skill and luck, and the Lions have definitely benefitted from some amount of luck here.
Variance is a very real thing, and you can’t just parse your way through it to a predictable outcome. The Packers/Lions game is one of the hardest to predict this season and I would stay far away from it in terms of gambling, but I feel better about the Packers than I did a few days ago. For starters, I think the Packer pass defense really is good, and the Lions will have issues moving the ball through the air should they need to. While Detroit does have an outstanding line and good running backs should this turn into an NFC North slop fest, the Packers have that as well, and with Dean Lowry out of the mix, I’m more confident in the Packers getting a stop or two in the run game.
The Packer defense relies heavily on turnovers in their own right, and it’s no sure thing Detroit starts giving the ball away this week, but some level of regression is likely, and Alexander did get Goff last time. More than anything, the last Packers-Lions game was a weird disaster. It was David Bakhtiari’s first game back, and he was rotated in and out with Yosh Nijman, which was not ideal for either. This was the game where Gary was lost for the season, meaning they had to adjust on the fly to the loss of their best defensive player. Dean Lowry played 60% of snaps, and Amari Rodgers was still the primary return man. Randall Cobb missed the game with an injury, Romeo Doubs got hurt on the first play of the game, and Christian Watson wasn’t fully integrated into the offense yet.
Perhaps the single biggest factor impacting this game was just how grueling it was from a scheduling perspective. It’s no coincidence that the Packers’ longest losing streak of the season coincided with the fallout from the London trip against the Giants, where they travelled to London for a “home” game, flew back to Green Bay to play the Jets in an actual home game, and then played three consecutive road games against the Commanders, Bills (on Sunday Night Football) and finally, Detroit. I don’t think any other team endured more difficult scheduling in any 5-game stretch this season, and that, more than anything, makes a repeat performance from the Packers unlikely.
Football teams only control so much, and the Lions’ variance is a testament to that. The Packers lost to the Commanders under Taylor Heinicke. The Lions beat them under Carson Wentz. The Lions missed Baker Mayfield on the Panthers, instead getting a surprisingly resilient Sam Darnold. The Packers got to face a washed Baker on the Rams. Stuff happens.
Going through it all, I like the Packers’ chances outside, with all of their receivers, their stud running back, and outside of Gary, a healthy defense playing about as well as it can. They excel in pass defense lately, something that has stymied the Lions all season, and unlike last time, if this turns into a shootout, the Packers have the horses to exploit an underwhelming Detroit defense. The Lions are a fun, weird team, and I think they have a bright future. They keep themselves in every game, and usually don’t beat themselves. They won’t roll over like Minnesota, but the Packers should get it done.