2022 Week 8 Reflections
Welcome (back) to Throwthedamball!
If this is your first time reflecting with us, I'd recommend first reading the inaugural piece of this series to get a sense of what I'm trying to do in this space.
Marco Thought:
This chart, which captures QB efficiency from a clean pocket—one of the more stable QB attributes— has so much to unpack. There is Zach Wilson in a tier with Burrow and Allen, Tua absolutely dominating, and Rodgers and Fields occupying nearly the same spot. But that’s not what I want to focus on here.
What is most striking to me is the position occupied by Geno Smith, Andy Dalton, and Jacoby Brissett. Why? These are all veteran QBs with long track records that before the season would indicate they are backup-level QBs. Certainly, these are QBs for whom it would be shocking from them to play at a near-elite clip in stable metrics for any extended period of time. Can they all regress? Sure, but I wouldn’t bank on it.
But this, in my view, gets at a fundamental point in understanding football and becoming a better bettor: football is complex, not complicated. What's the difference? Complicated problems are quite difficult to solve initially but once solved it is done for good. Building a car or airplane is complicated, but once cracked we know how to do it. Complex problems don’t really have permanent solutions. In as much as we might want to make football a complicated problem: “need an elite QB,” or “a good QB on a rookie deal”, or “good QB getting points” the game, and most importantly, interactions are constantly changing and complex. We can’t apply necessarily the same logic as we did 5 or even 2 years ago. We need to adapt as the game evolves and the interactions change.
If Geno Smith and Andy Dalton are succeeding in what I would’ve considered poor circumstances, I can’t help but turn back to first principles. I’m asking questions like: how much do we really know about QB evaluation? How much is dependent on scheme and context? Does that vary from team to team and QB to QB? Does the running game influence QB play in ways we might be unable to currently capture? Are there ways we can, in the future, better predict these interactions that led to such shocking outcomes?
Bottom line: don’t think we can just blindly assume what mattered most to team or player evaluation or handicapping in 2017 or 2020 is the same as it is right now. It’s an abstract and high level view but I think a helpful way to frame and evaluate play or betting. At the end of the day, we need to work with the best evidence available and adapt and change with the game. Ok, rant over.
Speaking of looking at the available evidence, here is how each team has performed in our Drive Quality metric so far this season.
Just an absolute mess outside of the Bills and Chiefs.
Football is wide open this year. Massive uncertainty about who is good means there’s inefficient markets—markets also don’t know who is good!— and that means there’s looooots of money to be made sifting through the noise.
Let’s turn quickly to some team trends:
TEN:
Somehow the Titans again have a performing defense and a commanding lead in the AFC South. But there are some worrying signs for the defense. The team is 7th worst in EPA on early downs(more stable) and is essentially tied for best defense in NFL on late downs.
ARI:
In 2021, on throws more than 20 yards, no QB had a higher PFF grade than Kyler Murray.
In 2022? Kyler has the 4th worst PFF grade throwing deep, and his EPA is full 1 point worse from last season.
I don’t think Kyler forgot to throw his deep ball. In each of his first 3 seasons he was top 10 in deep ball efficiency, so I’m going to guess this year is the outlier and noise more than anything else. I think he will move closer to his career average as the season progresses and the Cardinals offense will improve as a result.
LAC:
Justin Herbert checked down on 7.5% of his passes last year, but this year that number has shot up to 12.6% which is the highest in the NFL. I’m not sure if that’s a function of the injury, or the lack of receivers open, or some other variable. Either way, it is doesn’t seem ideal for Herbert to continually be checking down.
Bets:
Really did not like that much on the board this week, placing only 2 bets.
AZ -5.5 +285:
This was the featured ‘leveraging tails’ piece at PFF and I’d encourage you to read the full process there. Here’s the gist of it:
Law of Increasing Returns: The returns would diminish if a team like the Kansas City Chiefs added a player like Hopkins, because there is only so much more they can improve. But for Arizona, a team that struggles so mightily on scripted plays, the law of increasing returns suggests Hopkins' return will critically enhance this offense. On scripted plays in 2021, Hopkins was No. 1 in PFF grade, and No. 1 in Yard Per Route Run (YPPR). These increasing returns, I don’t think, are being properly priced into the market.
Vikings, Cardinals in same tier: Despite Arizona's struggles on scripted plays, at least according to EPA, these teams are roughly the same fundamentally. The Cardinals have produced a slight edge on offense, and the Vikings a slight edge on defense.
Cardinals aren't who market thinks they are: The market, and the No. 16 rating according to Inpredict , is seemingly pricing in Arizona's production so far this season. But this is before accounting for the projected increase in efficiency on the scripted plays and, even generally, Hopkins’ contribution to the offense. I would suggest a mispricing on the Cardinals here.
Arizona has played the 11th-hardest slate of opponents, while Minnesota has faced the 7th-easiest schedule, per PFF Greenline: And that's with the Vikings facing 2 backup QBs and having only one win by more than a full score (4-0 in 1-score games). Their 5-1 record, when looking at their place on the above graph and their previous matchups, looks unimpressive and speaks to potential value.
MIA - 5.5 & Tua O 230 passing yds +230
I’ve spoken numerous times about how sensitive Miami is to matchups considering how much they rely on their WR’s.
This matchup hearkens back to something Arjun Menon and I wrote for PFF this summer about how man coverage, more so than zone, lets talent play out. In man coverage, great coverage will almost always defeat weak receiving, and great receiving will almost always defeat weak coverage. And it's not just that this relationship exists in theory, but we know quarterbacks are very likely to throw the ball to specific players on account of the talent disparity whereas the same is not true in zone.
There was a fundamental mismatch between the Lions CB’s and Dolphins WR’s.
Above is a metric SCOE that Arjun Menon and I developed to evaluate coverage play. The bottom 3 players on this chart were the Lions starting CB’s and all have been among the leagues worst.
We don’t need a graph like the above to demonstrate the obvious: but Tyreek Hill and Jaylen Waddle are elite and both are among the very best in the NFL defeating man coverage.
Bottom line: with such a huge mismatch between receivers and CB’s exacerbated by the man coverage the Lions played. If the Dolphins were going to cover as I expected, Tua would have at least a decent passing day, and the 230 passing yards allowed us to improve our odds by nearly 100 without materially changing the thesis or introducing much variance at all.
Live betting:
The truth is most of the bets I place these days are live. I will always recap process for bets I make before the games, but they are certainly more limited than recent years. Either way, you can watch me stream my live bets every Sunday with Sportfoliokings—its a blast, we talk process, we celebrate and we tilt. I strongly recommend tuning in.
That’s it for this week. As always, thanks for reading and we will be back in this space next week! For graphs and articles on my process before Sunday, be sure to give me a follow on twitter @throwthedamball.
Please subscribe and share this post if you enjoyed and if you think others might stand to benefit as we together learn and improve our betting process!