Stealing Signals, Week 15, Part 1
Assigning credit/blame, plus TNF, Saturday, and early Sunday games
Before we get going today, a reminder for those of you still reading weekly that this will be the final week of the Stealing Signals Monday and Tuesday recaps for the year. It was just not viable this year to do a Week 16 edition with Christmas Eve falling on a Sunday and Christmas Day on Monday, particularly with the NFL scheduling three games on Christmas Day. Most of you know I’m a father of two, and I think I do a good job of balancing my schedule, but the family obligations take precedent here.
I will be back later this week for Input Volatility, like always, and I’ll have an Input Volatility for next week as well. Just not the Monday and Tuesday recaps.
One of the hardest things for casual observers is to parse how to dole out credit or blame for various things. I struggle with this, too, and am trying to be better — I’ve made some excuses for Bryce Young because I felt confident coming into the year that he had one of the most poorly constructed receiving corps you’ll ever see. But I have intentionally watched him more closely in the second half of the season, and I don’t really like what I’m seeing. It’s possible two things are true: the situation is terrible and he’s also at heightened risk for being a bust. If that’s the case, it would be wrong of me to spend the whole offseason defending him on the basis of his circumstances.
But I see this a lot in the feedback I get. One example is the Zach Wilson stuff, where every time I tweet about him I hear back how it’s really the offensive line, and no QB could have been successful in this situation. Or Nathaniel Hackett. And yes, those elements are not great, but also I’d say I’m highly confident Zach Wilson is not good. I don’t believe a good OL would change his penchant for back-breaking mistakes — often when he holds the ball too long — or fix his inaccuracy issues. He does make some good throws, too. But even in the context of a bad situation, you can watch him play and see a bad quarterback.
This runs in the face of my defending Justin Fields in his rookie season, where I did say we couldn’t analyze him because of a similarly terrible combination of offensive line and scheme. But for Fields, it was his first year — we’ve seen Zach struggle for three years now — and there was also a specific note in my writing then that he had traits that would flash, including accuracy and especially to deep parts of the field. He also added dynamism as a runner. Perhaps most specifically, it always felt like Fields’ situation was emphasizing his worst qualities, particularly his habit of holding the ball too long. And so it was, in my opinion, fair to want and need to see him in a different setting, where those potentially fatal flaws maybe weren’t so highlighted (to identify if they were, in fact, fatal, and would always prevent his other skills from shining through).
I’ve talked before about this idea of multi-variability in football, and how it’s the complexity of the sport that makes it so fun to analyze. But that multi-variability means we have to be able to think in the gray area. Football fans are always so quick to jump all over analysts for perceived inconsistencies in logic. Stats are often cited, and poorly. The cases being made are typically outside the context of the stats; I talk all the time about how most takes just follow the past week’s (or past year’s) box scores, and every time a player succeeds there’s this rush to find the haters and doubters and say, “Hey, look, he had a game with those stats, too, just like that other player at his position you’re always defending! You’re biased if you don’t like this player the same!”
But there is often not one singular cause for an event in football, despite our human nature to desire a neat little box to put everything into. It’s perfectly OK to say, for example, that Brock Purdy is not the MVP. That’s an opinion, but I do think it’s pretty telling that the analysts I know with Purdy MVP tickets have been basically going on a campaign on social to argue he has to win it, and their arguments only base that fact on data and how his stats mean he has to get it. I saw yesterday a claim that no QB with his stats has ever not won it, which of course ignores era adjustments, among other things.
I think Purdy has been excellent this year, and I also think it’s perfectly reasonable to argue he’s more replaceable and less valuable than other similarly successful quarterbacks. If he became the best statistical QB to not win the award, we could be intellectually honest about it and discuss how that sort of makes sense given his offense is probably the best collection of talent since the Kurt Warner-era Rams (it’s not just the four superstar skill position guys, but one of the very best left tackles in football, plus probably the best play designer in the game calling the shots; I think I’d argue this team around Purdy is better than Warner’s Greatest Show on Turf teams, even adjusted for era and those things).
The Purdy point fits this because people with a vested interest in him winning are making the types of arguments that focus solely on the one factor — his statistical success — while ignoring the other things the market has been telling us. See, Purdy’s odds have lagged behind what his statistics and team success would suggest about his potential to win the award, and that’s because the market does famously factor in more than just one data point. You can’t beat a market-based game armed with one piece of analysis.
In a fantasy space where so much of the analysis is tied to what has happened in recent box scores, and trying to retrofit analyses to explain results already finalized — rather than focusing on the predictive side of this — there is a clear edge to merely considering the parts of the argument that the obvious answer is missing. Being contrarian, if you will. Not for contrarian’s sake, but because we talk about probabilistic thinking around here, and even when there’s a dominant hypothesis for why something has occurred, that’s maybe only like 65% or 70% the cause of an event. Analysis will treat it like it’s 100%. If you can think through the range of other factors that make up that other 30% to 35% — all of which are probably like 5%, in some cases — and consider whether they support or perhaps alter the dominant hypothesis, you can often find little edges.
The better example is when the hypothesis that becomes dominant is maybe only explaining like 35% or 40% of an event, but we just don’t have any other strong answers and all the other little factors are again just like 5% or 10% increments. In those instances, all those little 5% things adding up to 60% or 65% gets ignored, because the focus still lands on the 35% or 40% thing given that it’s so much heavier than any other single factor. This is a weird attempt to explain this phenomenon mathematically, but I hope I’ve at least vaguely gotten there. The idea is that being diligent to the entire probabilistic range is going to create opportunity in the current fantasy football landscape.
That’s one of the last major lessons I’ll leave you with for the 2023 season, though I’ll be back with plenty of content in the early months of 2024, and as I noted still have Input Volatility coming your way the next couple of weeks. The note is that there are edges to be found in thinking through the secondary reasons for things we observe. The prime hypothesis gets so widely repeated in our space, that there sometimes isn’t room for anything else. But always be willing to consider alternate explanations; everything we do here could be summed up in this concept of probabilistic thinking, and that concept very rarely aligns with the neat little boxes we put things in.
(You know what’s a great example of this? Falcons owner Arthur Blank reportedly saying Arthur Smith’s job is safe because he understands the cap hell the Falcons were in when Smith took over. But Smith has had three years of control, and you can still evaluate not just the teams he’s put out there, but the development of the players, and even things like how he manages the media. It may be fair that Smith stepped into a bad situation and deserved time to build out of that hole, but you can still evaluate the ways he’s actually accomplished any of a reasonable set of goals over three seasons, and whether the other elements of the job suggest he is the right man to continue in charge for the franchise’s next phase. Blank has appeared unable or unwilling to do that, but there is still potential for him to reevaluate if the Falcons close poorly.)
Let’s get to the games, and I have 11 of them for you today, so really do need to work quick if possible. Data is typically courtesy of NFL fastR via the awesome Sam Hoppen, but I will also pull from RotoViz apps, Pro Football Reference, PFF, Next Gen Stats, Fantasy Life, the Fantasy Points Data Suite, and I get my PROE numbers from the great Michael Leone of Establish The Run. Part 1 of Week 1 included a glossary of important statistics to know for Stealing Signals.
You can always find an audio version of the posts in the Substack app, as well as easier-to-see versions of the visuals at the main site, bengretch.substack.com.
Raiders 63, Chargers 21
Key Stat: Zamir White — 71% snaps, 59% routes, 20 touches, 4 HVTs
This was an absolute shellacking, with a 42-0 scoreline at the half and the Chargers giving up en route to a -14.1% PROE. For the Raiders, there was a perfect storm element where they were coming off a shutout loss, and were always going to push things if they got the chance to make a statement here, while the Chargers were down key players and knew their coach was a lame duck.
Davante Adams (12-8-101-1) had a strong day as Aidan O’Connell looked the part, with Jakobi Meyers (4-2-32-1, plus a passing TD) and Michael Mayer (5-4-39-1) both also finding paydirt, and rookie Tre Tucker (4-3-59-2) getting there twice. The Raiders only totaled 378 yards, but won the turnover battle 5-0, and short fields were a big reason for the massive scoreline. Four of the offense’s seven TD drives started in plus territory, and then the defense also had two return scores. I’d be most concerned about how the TDs elevated the stat lines for the ancillary pieces, and far more comfortable with Adams’ production as he dominated targets with a 33% share.
Zamir White (17-69-1, 4-3-16) was in a much more dominant lead back role in Josh Jacobs’ absence than I have suggested I thought he would be, and it’s a reminder to never be too confident in these things and be willing to take chances on uncertainty if you don’t have great alternatives. The score and game situation also helped his workload, but the 59% routes were a real surprise, as I’d thought Ameer Abdullah (6-32, 2-2-13) might be in that range instead. It’s possible that in a trail script it could be a little different, but if Jacobs misses more time, I think you have to like what you saw from White here, including both green zone touches for the team. Brandon Bolden (2-25-1) only played two offensive snaps but got a 26-yard TD run in there because it was that kind of game for Las Vegas.
There’s not much to say about the Chargers, although they did wind up gaining quite a few yards. Quentin Johnston (3-2-23-1) caught a late TD, while Gerald Everett (8-5-41) led the team in targets and Joshua Palmer (4-4-113-1) hit for a long score on a busted coverage in his return. No other WR or TE made a receiving impact.
Austin Ekeler (5-9, 6-4-29) definitely shared some earlier snaps, and played later than you’d expect, although it’s tough to parse what all that meant because the Chargers were way behind early in the game, and the late work was probably just wanting to get him some work. Isaiah Spiller (16-50) easily led the team in carries, while Joshua Kelley (5-22, 1-1-6) also mixed in.
Signal: Zamir White — 71% snaps, 59% routes, both RB green zone touches
Noise: Raiders — 63 points, 5 passing TDs (four TD drives started in plus territory in a game with five turnovers); Brandon Bolden — 26-yard TD run (only played 2 offensive snaps); Austin Ekeler — 39% snaps (there was definitely an early rotation, and he does seem to be giving up reps, but this game was a blowout so early that it’s tough to read into the usage too much)
Bengals 27, Vikings 24
Key Stat: Ty Chandler — 79% snaps, 44% routes, 26 touches, 4 HVTs
In what became a fantastic finish, we were reminded in this Bengals-Vikings’ battle of backup QBs that back-and-forth fourth quarters can be our friends. I was tracking a fantasy matchup with Ja’Marr Chase (4-4-64), Joe Mixon (10-47-1, 3-3-14), and T.J. Hockenson (7-6-63), and I’m pretty sure they were all under 10 PPR points into the fourth quarter, but all padded those numbers late. In another world, maybe Chase doesn’t make the 24-yard catch to convert third-and-21 early in the fourth quarter, and the Vikings get the ball back with a 17-10 lead and effectively run clock. My point is that type of thing happens in other games often, and the difference between a disappointing final line or one that does get there late is often these game environment outcomes.
Chase did get knocked out of this game a little early, so he didn’t get to realize all of the late action. Tee Higgins (8-4-61-2) had a strong game, including one of the more impressive touchdown catches you’ll see when he undercut a DB on the sideline to basically reverse intercept a pass, then reached back to get the ball over the line. Tanner Hudson (5-5-49) was the only other WR or TE with more than two catches, but Tyler Boyd (5-2-53) made a clutch catch in traffic in overtime to set up the game-winning field goal. Jake Browning struggled at times in this one, but did make enough plays in the end, and his 8.9 passing aDOT was far more vertical than his prior two games.
Mixon wound up with a solid line, but early on Chase Brown (7-23, 3-3-28) got a lot of work, including touches on the second and third snaps from scrimmage. Brown only played 19% of the snaps in the end, with Mixon at 69%, but Mixon had done very little until being relied on late. It’s at least a mild concern for his reliability that the team is trying to incorporate Brown’s clear playmaking into the normal flow of the offense. Brown is electric with the ball in his hands, and this duo is a little reminiscent of the Melvin Gordon-Austin Ekeler tandem late in Gordon’s tenure with the Chargers.
Ty Chandler (23-132-1, 4-3-25) had a massive day, and looked far better than anything else the Vikings have put out there at RB this year. His 79% snaps and 44% routes were great, but his efficiency and production were even better, and I’d suspect he continues to play in something of a 50/50 split when Alexander Mattison returns, if not just staying ahead of Mattison.
Justin Jefferson (10-7-84) got through the whole game, and played well, but his presence seemed to positively impact Jordan Addison (6-6-111-2) even more. T.J. Hockenson (7-6-63) had his typical solid line as well, and the rest of the team combined for just four more receptions. A concentrated three-man receiving corps is great to see, and Addison is certainly worth consideration going forward.
Signal: Ty Chandler — handled big load on 79% snaps, and was efficient (likely has earned at least 50/50 share when Alexander Mattison is back)
Noise: Joe Mixon — 14.1 PPR points (this is only a minor thing, and Mixon wound up playing a strong snap share, but he was in real jeopardy of a floor game until late, as Chase Brown got designed touches early and it’s clear the Bengals want to incorporate Brown more)
Colts 30, Steelers 13
Key Stat: Jaylen Warren — 69% snaps, 83% routes, 7 HVTs (season high)
The Colts pulled away from the Steelers despite losing both of their most important pieces to early injury exits. Michael Pittman (5-4-78) was on his way to another strong game, but ran routes on just 47% of dropbacks. His typical RPO usage first seemed to go to Alec Pierce (5-1-13) with an RPO catch in the immediate aftermath of Pittman’s exit, but then it was D.J. Montgomery (4-2-48-1) who quickly seemed to step into that role. Montgomery did have a couple of drops on his four targets, and the volume spread around more than usual. It didn’t spread to Josh Downs (3-3-19), though, whose fall into obscurity was never as clear as it was here, when the team would have needed him to step up but also didn’t seem all that interested in relying on him as a key piece. Mo Alie-Cox (3-2-21-1) led the TEs, who were in their typical rotation.
Zack Moss (4-13, 3-3-20-1) was the other key piece that left early, hurting his previously broken arm on a receiving touchdown after an awkward horse collar tackle put unnatural strain on his fall. Trey Sermon (17-88) wound up handling plenty of carries late, and leading the backfield in snaps, but Tyler Goodson (11-69, 2-2-10) showed way more burst. My guess would be a timeshare here based on what we saw in this game — a whopping 15 of Sermon’s 17 carries came from the final five plays of the third quarter through the fourth quarter (starting on a drive where the Colts took over up 24-13 late in the third). Sermon is the bigger, more straight-line runner of the two, and they clearly leaned on that to run clock late, but prior to the final few drives, Goodson had been seemingly more involved, getting a first-down carry to being three of the five drives immediately after Moss’s exit.
Jaylen Warren (10-40, 6-5-28) finally got into a big snap share again, notably running routes on 83% of dropbacks while Najee Harris (12-33) was essentially a TRAP back and ran routes on just 13%. That split helped Warren to a season-high 7 HVTs, including two green zone touches. He would have scored on one run had George Pickens (7-3-47) showed any effort blocking downfield, instead allowing his man a free shot on Warren inside the 3-yard line.
Pickens did rack up a bunch of air yards on his 7 targets, but was unable to bring in much actual yardage. Diontae Johnson (6-4-62-1) led the way among the receivers, with Allen Robinson (3-3-19) and Pat Freiermuth (4-3-16) both posting disappointing lines in a game where Mitchell Trubisky threw for just 169 yards in negative script. The Steelers announced a switch to Mason Rudolph as their Week 16 starter, which likely won’t improve things much but is still a move I very much support given how Trubisky has played. Lean into uncertainty.
Signal: Jaylen Warren — 83% routes, 7 HVTs (season high in HVTs, and first time we’ve seen real upside in his role in a while)
Noise: Trey Sermon — 43% snaps, 17 carries relative to Tyler Goodson’s 27% and 11 (15 of Sermon’s came on the Colts’ final three drives as straight handoffs meant to run clock, while Goodson got far more work across five drives from the point of Zack Moss’s injury until the double-digit lead was secured; if Moss and Jonathan Taylor both miss, I’d expect a timeshare and slightly prefer Goodson)
Lions 42, Broncos 17
Key Stat: Jameson Williams — season-high 72% routes, 7 targets, 113 air yards
After a bit of a slow start on both sides, Jared Goff got back into a rhythm at home in the dome, throwing for five touchdowns, all to the main three targets we have been on all year. I was hopeful this outing was great for Stealing Signals subs, as Amon-Ra St. Brown (9-7-112-1) and Sam LaPorta (6-5-56-3) both smashed, and Jahmyr Gibbs (11-100-1, 2-2-8-1) continued to look like a young Jamaal Charles. The Lions play indoors each of the next two weeks, though a road game in Dallas for the fantasy championship might be difficult. Still, I’m optimistic these guys will play a big part in deciding league titles.
Jameson Williams (7-4-47) also had a notable game, seeing a season-high 7 targets with 113 air yards on a season-high 72% routes. He got a little banged up late so keep an eye on his health status, but there’s some potential for splashy outcomes here down the stretch. It’s not really an advised play for the fantasy playoffs, but perhaps he’s a name to mix in for those of you doing NFL playoff drafts on Underdog. Josh Reynolds (3-2-41) posted his second straight game with a routes number under 60%, something he’d done just once previously.
The Broncos couldn’t get much going, with Javonte Williams (12-27, 2-2-(-7)) struggling on the ground and also giving up some passing-down work in the negative script. Samaje Perine (6-37, 1-1-11) had a few chunk plays, and Jaleel McLaughlin (2-2, 3-2-16) had a near-miss at the goal line on a reception, which was followed up by Javonte also having one, both of which were arguably challengeable if Sean Payton had the nerve.
Courtland Sutton (6-5-71) and Jerry Jeudy (7-3-74) each had decent days, but as usual the receiving upside in this offense is just limited. It was rotational WR Lil’Jordan Humphrey (4-3-16-1) who caught the TD in this one.
Signal: Jameson Williams — season-high 72% routes, 7 targets, 113 air yards (Josh Reynolds was under 60% routes for the second straight week, and just the third time all year)
Noise: Javonte Williams — 20 yards (role was fine in neutral script, 14 touches, just an inefficient game, including being stopped at the six-inch line, and then gave up some late work in trail script)
Panthers 9, Falcons 7
Key Stat: Falcons — (-19.4%) PROE, Panthers — (-17.8%) (two lowest in Week 15, through Sunday)
In a rainy game, and with a clear run funnel matchup, you knew Arthur Smith would keep the ball on the ground. What you didn’t know was Tyler Allgeier (14-45, 1-1-6) would get the start and 14 inefficient carries, while Bijan Robinson (7-11, 3-1-3) was the third Falcons’ RB to touch the ball. Cordarrelle Patterson (5-(-2), 3-1-2-1) mixed in early, and then also at the goal line, and Bijan got just 8 touches despite a -19.4% PROE in a massively run=heavy game, in a loss to the one-win Panthers. No one out-levels himself quite like Artie.
Desmond Ridder completed just 12 passes in the conditions, with Kyle Pitts (4-3-37) and Drake London (3-2-24) joining Jonnu Smith (2-2-61) as the only guys catching multiple balls.
Bryce Young was better this week, completing 18-of-24 passes (75%) for 167 yards, a 7.0 YPA figure that was his second highest of the year. He did this while the passing aDOT stayed pretty high at 8.3 — that figure was not quite as high as the past two weeks, but still higher than all but one of his first 10 games (all under Frank Reich). Given the conditions and everything else, it was a minor victory in an otherwise really tough rookie season.
Adam Thielen (7-4-43) got back to leading the team in volume, but Jonathan Mingo (5-4-32) wasn’t far off. No one else on the team saw more than two targets, and the only other particularly notable item was Tommy Tremble (2-2-32) getting up to a 71% routes share, a season high for him as things consolidated for the Panthers.
Chuba Hubbard (22-87, 2-2-16) was the focal point of the offense in the run-heavy gameplan, and he continues to dominate the work over Miles Sanders (6-2, 2-2-11). Hubbard also just looks much better when you watch these two run.
Signal: Tommy Tremble — 71% routes (season high; all Panthers’ routes consolidated)
Noise: Falcons, Panthers — huge run leans (heavy rain, matchup)
Browns 20, Bears 17
Key Stat: David Njoku — 14 targets, 37% TPRR
In another rainy game, the Browns and Bears played slow early, but things did pick up late, particularly on the Browns’ side as they came from behind for the win, despite the Bears’ last-ditch efforts on a Hail Mary that was nearly completed.
David Njoku (14-10-104-1) continues to dominate, leading the way in a passing game led by Joe Flacco that hit a +6.0% PROE (third highest in Week 15, through Sunday) and totaled 45 pass attempts. Amari Cooper (8-4-109-1) also got loose for a game-tying touchdown late, as the added volume is paying dividends with Flacco under center. Njoku’s TD was a legit nice play in the back of the end zone, though I wouldn’t have been shocked if that one would have been overturned as it did appear his heel was on the back white line.
Elijah Moore (3-2-17) was quiet in this one, as rookie Cedric Tillman (8-4-52) was the third most notable piece, and it was veteran Marquise Goodwin (1-1-57) who hit for the the other big play on a deep shot. Tillman’s routes have been strong since the trade of Donovan Peoples-Jones, so this 8-target game isn’t really all that notable in the context of him posting just a 12% TPRR for the season so far, on a pretty solid 275-route sample (and also because 8 targets in a 45-attempt game is just a 17.7% target share, which isn’t exciting as a spike game).
Jerome Ford (8-20, 5-4-11) had a quiet game, and Kareem Hunt (7-8, 1-1-12) was worse, as the Browns’ heavy pass lean in the negative game script — and then Njoku’s role at his low 5.1-yard aDOT — both took from the RB volume.
Justin Fields (7-30) got just 5 designed carries in this one, easily his fewest since returning from injury, in a surprising development given the way the Browns defend the pass. D.J. Moore (8-4-52) played through injury to post a solid game, while Cole Kmet (7-5-23-1) found paydirt. Darnell Mooney (8-2-14) struggled on his volume, and rookie Tyler Scott (4-3-49) made a nice late play to nearly help set up a game-tying field goal attempt, before the team needed to settle for the Hail Mary.
We did get more information on the Bears’ RB situation, as Khalil Herbert (6-8, 1-1-4) continued to play a real limited role, and I’m content buying that now as a pretty clear indication he just isn’t back to 100% from his high-ankle injury, given how commonly that one lingers. D’Onta Foreman (6-(-6)) struggled in this one as well, and Roschon Johnson (5-36, 6-4-24) led the way with a 48% snap share and 42% routes despite what was mostly a positive script. I’ve said I didn’t think this three-way committee would play out in a way where I felt confident in anything by Week 17, but I do think it’s at least trended toward having some faith in Roschon.
Signal: David Njoku — 14 targets, 37% TPRR (continues to dominate volume, and also produce efficiently on it); Khalil Herbert — 24% snaps (doesn’t seem right in return from high-ankle injury); Roschon Johnson — 48% snaps, 42% routes in mostly positive script (has tended to play his biggest roles when trailing, so this seems at least mildly notable)
Noise: Cedric Tillman — 8 targets (17.7% share, 12% TPRR for the season on 275 routes, not actually impressive volume for a spike game given context that his role has been good for a couple months now)