This was my favorite tweet of the week.
Fantasy football is always chaotic, it’s always highly variable. We look for trends because that’s the lesser of two evils; the trends often don’t materialize, but waiting to see in every situation creates a roster stagnation that is as sure a way to lose your league as it gets.
We’ve been in the general range of the season where I’ve had the itch to write about the 2022 theme. I wrote about the “haves and have nots” last week, but one of the other big themes is this variability. Best ball has become a super popular game, and for good reason. It’s Week 5 and I’m ready to mention my long-running suggestion of having some sort of best ball element even in managed leagues — my favorite application of that idea was having two “bench” players that you still needed to identify in advance, much like any of your starters, but where they would replace one of your starters if they had a higher score. You could think about this differently as a league with deeper lineups where you drop the two lowest Flex scores (although I definitely think you should also get a backup QB, after the way QB injuries have hammered my lineups early this season).
I have two home leagues — one from my youth that’s been around for more than 20 years, and another with my college buddies that’s been around at least 15 or so. I probably take losses in these leagues harder than my many high stakes teams, which is a bit weird, not just because of the financial investment but also because I spend far less time on them, firing off waivers quickly and setting lineups pretty last minute. It’s I guess just an ego thing where they’ve been around forever and I know I’ll get a hard time from buddies whenever I have a tough season in a league like that.
And in those two home leagues, I can’t seem to keep a starting lineup healthy. In one of those leagues, this week was the second time through five weeks where I have had three separate starters leave the game early to injury. In the other league, I started Teddy Bridgewater — who scored literally zero point zero at the high-scoring QB position — and also Damien Harris, who exited a plum matchup early and watched his backfield mate go off for 161 rushing yards. It’s exceptionally hard to win fantasy matchups with outcomes like that. There’s just no way around that.
Anyway, I’m 1-4 in both, and some of that is my teams just aren’t very good, missing on some key decisions like riding with Kyle Pitts over Mark Andrews in one, and having Trey Lance in both. (The Pitts over Andrews thing has been the single most frustrating outcome of the early part of the season across my personal rosters, because I made a very specific point several times through draft season that I didn’t want my Pitts enthusiasm to keep me from getting Andrews exposure, even comparing it to how my DJ Moore enthusiasm last year kept me from getting more Cooper Kupp, and Kupp last year was for me a player I was very positive about and had several readers reach out to say my praise for him was why they wound up drafting him, though I didn’t get him on a ton of my own rosters. I recognized that specific risk of being too in “my guy” and how that would overlap negatively on my exposure of a player who could be this year’s Kupp! I don’t know all the places I said or wrote that, but in my TE Tiers piece — which is the only place I checked today because it seemed like where I would have made this point — I articulated it like this: “I’ve talked a ton about Pitts, but the main thing I’ve emphasized is I still want exposure to Andrews across my many rosters.” We are of course experiencing more or less the exact same outcome as Moore vs. Kupp circa 2021, in terms of me winding up with significantly more of “my guy,” Pitts hitting a similar outcome to Moore, and Andrews going on to be perhaps this year’s most important player. It’s been let’s just say not great for my sanity.)
This wasn’t necessarily the angle I was trying to go down with this introduction, but more to say that even among healthy rosters that are well-built, starting lineup decisions have felt trickier than ever before. I’m not sure if I’ve harped on this enough for it to be a well-known take, but those of you who follow my work closely will probably recognize it — I’ve talked over the past few years about macro trends that include teams getting deeper at positions like WR, getting more multiple and creative with how they run their offenses, and having a better understanding of long-term player health in a way that influences short-term decisions to err on the side of caution, and how all of those types of things are changing the fantasy landscape.
These are all good things, especially the player health element, but even the coaches who recognize that to win NFL games they need to vary their approach week in and week out and be one step ahead of what they are putting on film and expect the defense to respond to. Football has long been about execution — there are so many teams throughout NFL history that got very good at specific things and did them so well that they could be successful in the face of added defensive attention on those things. That focus on execution will never change, but what we’re seeing in an era of technological advancement is coaches and teams recognizing their opponents will have certain information, and deception and variability in play design and approach are easy buttons for that execution. Again, this has also been around for a long time; my suggestion is merely that as time passes, offensive coaches are generally more willing to try to stay one step ahead than try to perfect a specific scheme. There’s always been self scouting, but it feels like we see teams very aware of what they are putting on film, and how defenses might approach them, and then they proactively build off that in ways that are meant to anticipate their opponents’ actions.
The weekly variability of certain types of football players’ stats has been long established; I’m constantly talking about it. Again, all I’m suggesting is things are shifting toward even more variability, and even less predictability for us. This is sort of rich coming off a week in the DFS world where the chalkiest plays all seemed to hit, but it’s also well-established over there that the chalky plays these days are often not the players who have succeeded recently, but those who have failed and gotten cheaper and are probably due to succeed soon. I would also argue there are still situations we can trust to be concentrated — I’m not saying every single offense is highly variable — and so maybe it’s not as much about the obvious “chalk” players but then the pivots off those players, those second and third tier players looking great one week and then bad the next, making things difficult.
One main thrust of this introduction was meant to comment on how reactive we are.
I saw this tweet and then saw a reply that said something like “they don’t trust him to throw,” which if all you knew about Jalen Hurts was what he did in Weeks 2 and 3, that would seem like an insane opinion. Prior to MNF, he’s sixth in the NFL in passing yardage on a team that is 5-0 and thus hasn’t had many situations where they’ve been forced to the pass! Of course, if all you knew about him was the narrative surrounding him all offseason and his career before 2022, and then you saw Weeks 4 and 5, you would probably be led to that opinion. The point isn’t to make a whole thing about one random Twitter comment, but rather to discuss something I have said before, that the great Fantasy Douche used to repeat, which is that a really high percentage of football opinions can probably be explained by just looking at last week’s box score. We can expand that a bit as more and more data gets widely accepted and these narratives get a bit more nuanced to say “looking at last week’s data,” and it’s funny because I think with more nuanced data we get tricked even further into thinking the trends must be believable. When you see something like this Next Gen Stats data, it can feel even more powerful toward the confirmation bias of the idea Hurts isn’t being trusted.
Except that’s not really the case at all, right? We have five weeks of data, and especially in Weeks 2 and 3, the Eagles really pushed the ball downfield, with Hurts registering passing aDOTs those weeks of 9.3 and 11.3 yards, and throwing for over 300 yards in both games. My interpretation of this data from Hurts this week is just that the Eagles had a different gameplan, and knew they could and maybe should mix it up, just like they did when they went run-heavy in a game with some rain in Week 4 and decided to try to win that game differently than they had in Weeks 2 and 3. They’ve now attacked defenses with a run-heaviness, with downfield passing, and with a high dosage of the short passing game (in both Weeks 1 and 5). It’s probably a big reason they have been so effective on offense and sit at 5-0. They weren’t perfect this week, but they got by, somewhat due to Arizona’s late-game errors.
By not being predictable, they are staying one step ahead, and they are also pretty maddening on a week to week fantasy basis in terms of how much we should trust their players. Last week, Miles Sanders looked amazing. This week, DeVonta Smith looks like the higher-volume guy than A.J. Brown, and Dallas Goedert is coming off his best game of the year. We’re actually blessed that this is a very concentrated offense, as far as these things go, but even within that concentration the narratives at any point on these specific players probably just followed the box score from the previous week.
When I started writing this intro, I didn’t even have the Eagles in mind. I’d forgotten I’d set that Hurts tweet aside, and even that was just going to be a point about how he’s thrown downfield in other weeks, and how his passing narrative has shifted. But as his passing narrative has shifted, so too have the narratives of the skill players on his offense, and this is on perhaps the league’s most surprising offense overall, one where everyone feels like a hit. What I’m saying is we could do similar with basically any other situation — Gabe Davis’ role with the Bills looks a lot more like his preseason expectations today then it did on Saturday, Breece Hall looks like the absolute superstar we thought he could be, concern about Austin Ekeler and Leonard Fournette’s HVTs was obviously overstated, Taysom Hill was always due for a game like this, doubting David Njoku’s consistency was a mistake, Matt Rhule was obviously going to get fired, Rashaad Penny was obviously going to get hurt. I made that list in two minutes while mostly just looking at the top of this week’s fantasy producers, i.e. driving my narratives from the box score results.
So what does any of this mean in terms of actionability? I think now more than ever we need to stay focused on the “long view.” I’ve heard from a few of you that were grateful for my comments on Hall over the past few weeks, which is kind of funny because in last week’s Input Volatility piece I brought up some potential concerns as well. It’s sort of all about the angle you’re approaching it from, and I did try to keep that commentary optimistic overall. But I mean, why couldn’t Hall’s big Week 5 have been Garrett Wilson? It very obviously could have, right? But we’re instead very down on Wilson right now, collectively, because Zach Wilson took over the offense and the worst possible outcome for the young WRs has transpired through two weeks there.
Is that a trend? Is that Signal or Noise? It feels like Signal to me, but that’s probably the recency bias talking. I still think Garrett Wilson is an extremely good player, and Elijah Moore producing at this level all season seems impossible to believe,
But a criticism I’ve heard about this column that I’ve mentioned before is that I stick too heavily to my priors. It’s one of the things that has me questioning the actionability of the game-by-game stuff at all, and why my approach with Stealing Signals has shifted. I think in a lot of these situations, the correct approach is to do nothing. To be shifting your opinions more slowly than the market. That’s not satisfying at all. It’s also probably wrong to do in every situation, because that bias for action on waivers and in trades is an asset overall.
It’s all a bit difficult to keep straight right now, if I’m honest, which I always am, even when my thoughts aren’t fully formulated like they aren’t this morning. But right now it feels to me like a lot of the week-to-week, short-term analyses in the industry are more or less just guesses at ways to interpret data that could be meaningful, but where you could make the exact opposite argument, that it’s an overcorrection. That’s not to say we shouldn’t have any approach, and things have always been able to shift pretty quickly. But the other main thrust of this introduction was to make a case those weekly value shifts are more prevalent in 2022 than ever before.
Everything looks clear in hindsight. Or, I don’t know, at least to some people. I guess for me, nothing looks clear in hindsight. All I see is a bunch of data that tells me the next places we’re going aren’t going to be predictable. And I’m out here writing about Signal and Noise every week! I do have to acknowledge this isn’t a universal truth by any means — I feel strongly that Mark Andrews is going to keep crushing, for example — but I have no idea whether Rhule’s firing is going to unlock DJM (seems unlikely, right? But that’s what recent box scores would suggest), or whether Gabe Davis’ Week 5 is indicative of his upside in this offense or further concern about his target share, or which player might get injured next and how those teams might respond, which to be fair is always a concern.
I’ve always been a bit weaker with my convictions than some, and I tend to wear my emotions on my sleeve throughout the season. This is probably just a week where I’m a little less optimistic about our ability to read a lot of stuff, for whatever reason. There have been some serious injuries over the past couple weeks, we’ve had a few things swing back pretty aggressively. Maybe that’s all these musings are today, is more bias toward recent outcomes. (Or maybe they are just me annoyed the NFL can’t balance the schedule and I have 11 games to write, so I’m just procrastinating with this intro. The bye weeks are coming next week!)
The further we get into the season, with more data, there will be clearer pictures. That’s always the case. But as far as how I’m playing my 1-4 teams, for example — I’m just hoping I’ll get through a week where all the players I put in the starting lineup finish the game. It’s pretty funny to look at my 5-0 and 4-1 teams — it’s not even that I don’t think there’s a meaningful difference between those rosters and my 1-4 teams, it’s that in a couple cases I think those teams are worse.
Now that I’ve driven home my apathy, let’s get to the Week 5 games. There are 11 today, and I’ve written a couple thousand words this morning about our inability to predict anything, so let’s have a little fun with it.
Data is typically courtesy of NFL fastR via the awesome Sam Hoppen, but I also pull from RotoViz apps, Pro Football Reference, PFF, RotoGrinders, Add More Funds, 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.
Keep an eye out for the audio version of Stealing Signals, found in the Substack Reader app.
Colts 12, Broncos 9
RB Snap Notes: Deon Jackson: 58% (+54 vs. season high), Phillip Lindsay: 38% (debut), Nyheim Hines: 4% (concussion), Melvin Gordon: 56% (+15 vs. high), Mike Boone: 41% (+5 vs. Week 4 high)
WR Snap Notes: Alec Pierce: 59% (+8 vs. high), KJ Hamler: 53% (highest since Week 1)
TE Snap Notes: Mo Alie-Cox: 44% (-26 vs. W4 high), Eric Saubert: 56% (-21 vs. W4 high)
Key Stat: Russell Wilson — 10-for-29 on passes traveling 20+ air yards (would be worst deep completion percentage of his career)
Could there be a better first game than this to support my intro? Does this even qualify as a football game? It was more of a fever dream. The players and the coaches acknowledged as much after the game. Am I supposed to write about whether I think Russell Wilson literally forgot how to play football? Or whether this whole thing about a lat injury in Week 4 explains his terrible play since Week 1? I need an expertly-produced five-hour documentary on that subject. And we’ll need that ready this week. (That would do numbers, because it would have the exact thing you need for a documentary to do numbers, which is a whole bunch of theories but no real answers.)
The Colts look bad in the more traditional way. Where the Broncos had opportunities — and a quarterback with arm strength to get the ball there who missed the spot to throw to by a full 10 yards in a way that made me question whether a) he really is that inaccurate now, b) he somehow didn’t see/thought there was a better spot to throw to, or c) everything I know about football is wrong and he’s so unlimited he’s ascended to a different plane where he can see things I can’t even imagine seeing — the Colts have a quarterback who doesn’t have the arm strength to throw the ball to the sideline on a 10-yard out pattern.
Michael Pittman (8-5-59) had another game where his QB and situation limited his effectiveness. Alec Pierce (9-8-81) had something of a breakout game volume-wise, and with three straight contests putting up at least 60 yards, he’s starting to look like a solid No. 2, not in a way that matters much for fantasy but in a way that might matter (negatively) for Pittman. Pierce’s 77% routes were a season high.
Mo Alie-Cox (1-1-7) turned into a pumpkin, seeing his routes fall all the way back to 38%. Until the Colts look consistently capable of scoring offensive touchdowns, which they haven’t yet this season, you can safely ignore their passing weapons outside Pittman and maybe stashing Pierce.
Nyheim Hines (1-3, 2-1-5) unfortunately suffered a concussion on the third play from scrimmage, all three of which were plays that went to him. It looked like his opportunity to get a lot of run, and instead Deon Jackson (13-62, 4-4-29) and Phillip Lindsay (11-40, 4-3-14) split the work pretty evenly, which Jackson leading in routes at 53%.
Melvin Gordon (15-54, 3-3-49) led the Broncos backfield, with Mike Boone (7-38, 3-3-47) playing a solid amount and showing some explosiveness, which he’s known for. The routes were interestingly split, with Gordon at 47% and Boone at 42%, and Gordon got both green zone touches. We’ll need to see how Latavius Murray fits in next week after he was inactive for this Thursday nighter due to having played with the Saints in London the prior Sunday. This was a great outcome for the idea Boone could hold off Murray and keep this as a two-back committee, or potentially even threaten Gordon’s work. It wasn’t a great sign for Gordon, though the green zone work was something.
Courtland Sutton (11-5-74) continues to be Wilson’s favorite target downfield, and his 148 air yards were second most in Week 5, through Sunday. Jerry Jeudy (8-3-53) had 109 air yards of his own and several near misses, including one in the back of the end zone on an uncharacteristically well-thrown ball that Jeudy seemed to lack effort on, trying to catch it with one hand while maybe being held, but doing nothing to draw the flag. Later, he was open on a wheel route where Wilson just overthrew him for what should have been a TD. The play that frustrated me most was when Jeudy got wide open and had an easy big gain if Wilson just puts this ball between the numbers and the sideline, somewhere between the 45s, rather than chucking it to the other 40-yard line and back toward the field, letting the defensive back get back into the play.
It’s such a bad throw because there’s no defender underneath the route that’s in the play, so like I said Wilson could have put this throw to their own 45 or the 50, and Jeudy could have flattened his route with the ball in the air, caught it, and probably still turned and had space to make a move after the catch. Some QBs are very good at understanding where the miss is relative to where the defenders are, and just securing this big play with a very simple “underthrown” completion where Jeudy could have more or less slowed to a stop and caught this ball (but likely not picked up yards after the catch) would have been much preferred to the actual outcome. What you can’t do is miss exactly where Wilson did, upfield and away from the sideline, running Jeudy back into the only defender who had even a vague shot to impact the play at the catch point.
KJ Hamler (2-1-10) got back to 71% routes, a big increase for him and a positive sign, though he basically didn’t get looked at, including on the final play where he could have walked in for the game-winning touchdown, which he correctly noted after the game. Hamler’s a dude I’ve liked, and in the spirit of the intro, I can still see outcomes where he’s relevant if this passing game figures their issues out, but there’s just no real reason for optimism on that front. Eric Saubert (7-5-36) continued to look like the lead TE, but his routes fell dramatically from 87% to 53% of dropbacks.
Signal: Alec Pierce — season-high 77% routes, 9 targets, now with three straight games over 60 yards (not startable in this offense right now, though he looks like a bit of a problem for Michael Pittman); Mike Boone — 41% snaps, 42% routes, showed some explosiveness (strong outcome for him as he hopes to consolidate work alongside Melvin Gordon with Latavius Murray looming); Melvin Gordon — both RB green zone touches
Noise: Eric Saubert — 7 targets (routes on just 53% of dropbacks, down substantially from his 87% routes in Week 4); Jerry Jeudy — 3 catches on 8 targets (was very open for multiple incompletions on potential big plays, and maybe that’s not Noise at this point, but I have to imagine there will be connections eventually)