Stealing Signals, Week 6, Part 1
Being in and out of position analytically, plus TNF and early Sunday games
Man, what a week. There’s just so much to discuss, and here I am still hazy from a Saturday trip to the Washington-Oregon game where I not only saw my university win a huge game against a hated rival, but do so after hope seemed lost late.
(For anyone who watched that game, and there’s been much ado over fourth-down decisions, I can’t really describe how demoralizing Bucky Irving’s second-down run was, which immediately preceded the Huskies getting the massive third- and fourth-and-short stops, where a conversion could have ended the game. In the lead-up to Irving’s run, Washington had been stopped at the goal line on three consecutive plays from the 1- or 2-yard line — including two shotgun runs from the 1 that drove me nuts — and then Oregon had driven back to around midfield from the shadow of their own goal line, with a four-point lead and a chance to run out the clock, having converted two third-and-shorts. And that was it — we absolutely had to have the stop at that point, and couldn’t give up another first down. We stuffed them on first-and-10, and then seemingly had Irving corralled on second down as well, for what might have generated a crucial third-and-long, but then Irving slipped away from multiple unblocked defenders around the line of scrimmage to instead gain eight yards on a play that looked like it was going to generate at most two. And it felt like that was it, we weren’t going to get the stop we needed. It’s still so insane that four plays later Washington had the lead. And then a few minutes after that, Oregon missed a potential overtime-inducing field goal as time expired and Washington had won. That Irving run was so demoralizing, but sometimes football is just like that; we did get the third- and fourth-and-short stops, then also scored in two plays, with completions of 35 and 18 yards, and then that held up. Just madness. I have most definitely not recovered from that whirlwind of emotions.)
Anyway, Sunday was crazy too, and the Monday morning discussion is about how low-scoring it all was, which I certainly noticed. During the games, I went and checked where we were entering the week, and I was surprised to see we hadn’t risen back that much from 2022. I’ve linked this Pro-Football-Reference page before, and it’s one I head to frequently, as it shows per-team-game averages of a variety of stats for every season in NFL history.
PFR updates that stuff overnight, or early the following morning, but not in real time. So when I was checking on the stats during the games yesterday, it reflected everything before yesterday. I was frankly surprised to see the average points for at 22.2, compared to 21.9 last year and 23.0 in 2021. Now — after yesterday’s debacle — we lost another half point and are down to 21.7 points per team game, fewer than last year. We’re down to 5.2 yards per play, fewest in over 15 years. We’re down to 5.9 NY/A, the lowest passing efficiency in almost 20 years, but while rushing yards per carry reached an NFL record 4.5 last season, to somewhat offset the lack of explosive passing, we’ve also lost the rushing efficiency here in 2023, as it has fallen back down to 4.2 so far.
I don’t bring this up to draw any conclusions. It’s a small sample, and I’ve been chasing my tail on the macro stuff a little. In fact, that’s what I want to talk about. This idea of being out of position analytically to draw clean takeaways.
Over the years, I’ve spent a lot of time in these pages not just analyzing football, but analyzing the ways that we analyze. One of the funny things when you write about biases is you’ll feel pretty strongly something has been written about before, and been discussed, and has a name, but also that it isn’t easy to find. There are just so many biases out there — I’ve spent hours of my life on the Wikipedia page, “List of cognitive biases,” which features sections and subsections and tables, and just dozens and dozens of biases whose meanings in some cases feel overlapping, or like they could work in concert with one another — that it’s not exactly easy to search out some specific thing that’s frankly not easy to explain anyway, because we’re talking about unconscious ways our thinking is influenced.
Anyway, I’ve been thinking a lot lately about how analysis is influenced by past analysis, and I’m becoming increasingly convinced our decisions are subtly more tied together than it might immediate seem. I mean, there are ways things tie together over time that we’re all pretty aware of, relating to the widely-discussed biases like confirmation bias or recency bias, or when we make arguments in favor of players and then anchor to those arguments and defend them in the face of new evidence. All of those things play into what I want to articulate this morning.
But it’s a slightly broader feeling. It’s this idea that when you’re right, and specifically right for the right reasons, and you identified something in advance, that you’re better positioned to analyze the next trend. And that also might seem obvious — we’re always analyzing the market, and so if you recognize a mispriced element, it makes sense you’d be better positioned to identify any overreactions, or the next mispricing.
I always thought it was fascinating, when I would follow DFS a little more closely, that really good players seemed to go on “heaters.” Theoretically, every week is a unique puzzle, with new pricing, and new strategies, and the contests are independent of each other. But the case I’d make here is when you identify some key breakout elements one week — whether it be a player or an offense or something even broader, like a positional trend — you go into the next week, and the new price atmosphere, with a better foundation of what things really are.
The opposite of this might be a fixation on what you see as a mispricing, but that isn’t hitting, e.g. chasing Kyle Pitts week in and week out over the past season and a half. The person fading Pitts might be better positioned to consider when to get back in and play him, at a time when ownership falls to a very low point due to an overreaction. More accurately, the person fading Pitts isn’t that worried about Pitts. Their focus isn’t on whether the results and process were misaligned because of variance or because the process was wrong. That person’s process aligning with the results allows them to pivot to the next piece of forward-looking analysis, rather than chasing.
I’m not sure if I’m making this point as deftly as I’d like. (You may be understanding why I opened this intro talking about still being in a fog from Saturday.) But I’ve been thinking about this bias, and how the conclusions we draw seem to be influenced by how accurate our past conclusions were, and how some people really seem to struggle with that. I bring it up in relation to the lack of scoring, because there’s this minor urge for me to be like “See, this is what I was talking about!” in relation to my past writing on these league-wide trends. But I can’t really do that — after Week 2, I was thrilled to see the explosives back, and I felt my prior analysis was wrong, and I accepted that and reacted to the new normal.
Again I fear I’m not making this point with the nuance I want. But my point is not to do some humblebrag, or beat myself up. It’s specifically to say that in the time since I celebrated the return of offense in Week 2, I have made faulty assumptions about a trend I was late to, which wound up not being a trend at all. The idea here is me realizing that by immediately capitulating — something I’ve written about before that everyone seems to want in the fantasy industry these days — I was not only late to something, but I was massively out of position when it became plausible that my initial position perhaps did have some accuracy to it.
One of the ways this has manifested is with over/under picks I’ve made over at Stealing Lines. After the scoring issues in Week 1, I took a bunch of unders in Week 2, and then overs went like 12-4 that week and I got wrecked. And I legitimately bought into Week 2 as the rebound, so in the time since, I’ve had a bias toward picking overs, as I’ve seen the scoring environment more favorably. But I was late to that trend, such that it was a thing at all — people getting into overs for Week 2, after the average game total dropped by almost two full points from Week 1 to Week 2, would have been ahead of the curve, and then perhaps able to recognize when it was time to get back onto a bias to the under.
I know this because I’ve been there before — I recognized the elevated scoring environment early in 2020, and did very well betting overs that year. And last year, I pivoted to a bias to unders at an early juncture to some success (although the market always pivots, as well, and over/unders fell quickly last year; the key is you have to be ahead of the trend and not behind it).
This isn’t to say I should have dug my heels in on my initial analysis, or even that my initial analysis was correct, because at this point I do not have a high degree of confidence in where we’re at with the scoring environment. We have some real QB issues around the league. I do think some scheme things have shifted, where it’s felt like offenses have been able to adjust to the rolling coverages and post-snap defensive changes, but I don’t really know the specifics well enough to comment, and it’s clear in some cases the answers aren’t all there (Derek Carr for example highlighted miscommunication in reacting to post-snap defensive adjustments for some of the Saints’ issues in his presser this week).
The broader point is a lesson in how reactionary analysis often makes you wrong regardless. What I don’t want to do now is start hammering unders again, like I did in Week 2. To tie this back to fantasy, one of the biggest notes I’ve seen in this stage of the season is how people want to cut players that are burning a hole on their bench, and I see the cuts for players there’s no justification to add over them. It’s purely an emotional response to wanting to get rid of someone rather than needing to add someone else.
That’s not to say you hold players forever, and there are always spots that could go either way. But the point is that sticking with players in deeper leagues despite negative trends — and the emotional feeling of being wrong on them — can still turn around and eventually hit. I heard from a few of you that really liked my “stash Michael Mayer” call last week, and that’s a dude that looked pretty dead in the water after a couple weeks, and then after showing some life last week, this week he consolidated the TE snaps up to 81%, caught 5-of-6 targets for 75 yards, and suddenly looks like one of the top waiver wire targets of the week in a concentrated offense with room for a rookie hit.
I bring up Mayer because basically anyone who drafted him had cut him long ago; his ownership even in deep TE Premium formats like FFPC was very low when I checked last week. And yet if even someone like that can see the upside scenario materialize seemingly out of nowhere, then moving on from an even more popular stash just to move on — that’s risking making you wrong twice, like I described with the scoring environment earlier. (That said, even if you had moved on from someone like Mayer, you could do so in a way where you’re not convincing yourself he definitely sucks, and are instead still very open to the early signs his upside scenario is hitting, such that you can react in blind bidding this week in a way that gives you the best chance of adding him.)
I guess the other guy to discuss in this context is Adam Thielen, who posted another fantastic performance. Thielen looks awesome, and every year there are seemingly one or two veterans like him who perform very well for stretches, and then the focus seems to be on getting everyone who has ever disliked older players to admit they were wrong.
But as Nick Saban might say, “I’m not going to, so quit asking.”
Nah, I’m just kidding. I’ll readily admit I was wrong here. You have to when a dude scores as many points as Thielen has. But there’s also no world where his performance is actionable for me, because I’m going to be behind the market. I even wrote last week that “With any sell high, there’s an element where it’s possible what he’s doing continues in the short term,” which isn’t to play both sides of the coin as much as to say this isn’t actually one where I was shocked by his Week 6 line and am having to scramble to react. Friend of the newsletter Pat Kerrane referenced my thoughts on Thielen in his great Walkthrough writeup last week, agreed, and then talked about how this spot was the exact situation where he would keep up what he’s been doing, which he nailed.
But the point remains that over a longer timeline, Thielen’s production is fragile. The list of players who profile like him having really impressive stretches is a lot longer than the ones who did it for full seasons. I won’t try to analyze the specifics other than to vaguely say there are a lot of things that could happen; one possibility is defenses have been more or less willing to allow his specific type of production, but if teams start to see it as a bigger threat, more defensive attention could shut it down.
Has Thielen impressed? Of course. He has more juice left than I expected. But there are a lot of people that are quick to be a lot more excited about what he’s done than I will be. And I’m not really trying to make this about Thielen any more than I was trying to about the macro trends; the point remains that if you’re constantly reactionary, you’ll also be constantly late. Meanwhile, if you more or less ignore the situations you’re wrong on — like I have been with Thielen, or Week 2 being proof the scoring was back — and refuse to capitulate but also not necessarily dig your heels in, at least you will be in a unique lane, able to quickly take advantage if things shift back your direction. There’s an advantage to perceived stubbornness.
We’re all making a number of decisions every week. One of the things you guys have heard me complain about over the years is the focus on the negative, because I am very confident in the number of hits I have as well. And a big part of that is all tied up in this stuff I’m discussing in today’s intro. I’d much rather give my opinion on stuff I have been right about to this point, because that’s where I’m going to be most credible. I know in the past I’ve used a line like, “I’m not the person whose opinion on [Thielen] you should care about.” I would just be chasing something I didn’t believe in, but the goal here is predicting future outcomes, and the obsession with taking Ls puts you behind an analytical eight ball going forward.
But that’s where what might be the perceived stubbornness from me — that’s where it comes from. I’m not against admitting misses, but I’m also going to keep the focus forward, and specifically on situations I actually do have insight into. The huge point on a guy like Thielen is the value is rising, and people are increasingly buying in every week. Too many times I’ve responded in kind, and put myself out of position analytically — late to one trend, and blind to when it snaps back. I definitely need to react to some degree, and I thought for example playing Thielen in DFS this week was sharp, based on Pat’s analysis. But the focus should be the players I’m in position on, analytically, and the spots I have read well from the start. And that’s where I want to try to keep my focus week in and week out.
I didn’t write that intro as well as I wanted, but hopefully it made some sense. Let’s get to the games, which weren’t fun this week! A lot of the Key Stats this week are team-level stuff, because it just wasn’t a great week for big stat lines from various players.
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.
Chiefs 19, Broncos 8
Key Stat: Broncos — 117 air yards (fewest by any team in a game since Week 1)
Week 6 kicked off with a pretty boring game last Thursday. The Key Stat references Denver’s unwillingness to throw downfield, though Kansas City actually had a lower team aDOT at 4.2, and only 174 air yards of their own. Isiah Pacheco (16-62, 6-6-36) had a nice game, while the other two backs continued to rotate in, and then Travis Kelce (9-9-124) dominated the targets while the WRs continued to rotate.
Perhaps most notable to the rotation was Justin Watson’s (1-0-0) injury, which is likely to cost him some time. Rashee Rice (4-4-72) looked very good again, and could be the beneficiary of Watson’s injury. Rice’s routes were back up to 50%, matching Skyy Moore (4-2-22), whose routes continued to dip and were at a season low here. Marquez Valdes-Scantling (1-0-0) continues to run the most routes but do absolutely nothing, production-wise. Kadarius Toney (6-3-9-1, 2-(-5)) fell back a bit, but did score a short touchdown. Everything continues to point to Rice being the guy to stash, which hopefully we get a longer look with Watson out, and then the real window for that to pay dividends is likely after the Week 10 bye. Still, if you’re in a shallow league where he’s out there, this week is a great time to act on adding him. His per-route production remains very strong, and in this one he showed strong after-the-catch ability, turning just 17 total air yards into 72 receiving yards, which is just one more positive in his overall receiving profile.
Jaleel McLaughlin (7-30, 2-2-12) got the start and first touch of the game for Denver, while Javonte Williams (10-52) also got a lot of early work in his return from injury. In somewhat of a surprise after Samaje Perine (2-2-16) played 60% of the Week 5 snaps (ahead of McLaughlin), Perine was down to just 9 offensive snaps here (17%). If the rotation stays largely between McLaughlin and Williams, that could be pretty beneficial, although I remain skeptical of the whole offense. It’s also possible Perine made way for Williams because Williams hadn’t played Sunday, Perine had (and had played a lot), and it was a short turnaround for this Thursday night game.
Circling back to being skeptical of the whole offense, the Broncos totaled just 197 total yards here — including just 117 air yards and 22 pass attempts in a losing effort — and you have to be particularly skeptical of the receiving weapons as the offense continues to flow through the RBs, plus the edge defender moves (getting rid of both Randy Gregory and Frank Clark) and rumors of Jerry Jeudy (5-3-14) being available giving the impression Denver might be turning a page.
Courtland Sutton (6-4-46-1) was the main target, and other than Jeudy, the only other Bronco to have more than 10 receiving yards in a game where Russell Wilson threw for just 95 total (and took another four sacks). Greg Dulcich (2-1-3) made his return but had to leave due to injury again. Marvin Mims (no targets) is just a figment of our collective imagination (that you should still hold as long as you can, in case a Jeudy trade does go down, but in some shallower leagues I get needing the space as well, because we don’t even know what would happen to make this situation better post-Jeudy anyway).
Signal: Rashee Rice — back to 50% routes, could benefit from Justin Watson’s injury if WR rotation tightens up a bit; Broncos — 22 pass attempts, 117 air yards in a loss (really tough to trust their receiving weapons right now)
Noise: Broncos RBs — tough to read the trend, but moved to more of a two-back committee with Samaje Perine the odd-man out, which seems plausible to stick or potentially the result of Perine playing a lot on Sunday and it being a short week (feels like a favorable commentary on Jaleel McLaughlin either way)
Ravens 24, Titans 16
Key Stat: Titans — 46 plays, 21 pass attempts (both fewest in Week 6, through Sunday)
I suppose I could start each of these games this week by noting it was a pretty boring game, but Tennessee threw just 21 passes in a loss, while the Ravens — much like the Chiefs in the last game — coasted to a victory because their opponent wasn’t really pushing them.
I speculated last week the Ravens were treating the RB position as a hot hand situation, but that wasn’t true here, as Justice Hill (8-35, 3-3-0) pretty clearly looked better than Gus Edwards (16-41, 1-1-12), but the team pretty clearly prioritized Edwards as the preferred between-the-tackles runner to grind out yards against a tough run D, with the smaller Hill as a change-of-pace guy they tried to get on the edge with outside runs. I continue to remain optimistic for Hill’s ability to grow into a valuable role, but the usage here was a positive note in Edwards’ favor, in terms of how the team views him.
Zay Flowers (8-6-50-14) and Mark Andrews (6-4-69) led the passing game, while the other three WRs rotated. Nelson Agholor (3-2-40) wound up next in routes, while Odell Beckham (4-2-34) was an afterthought. Sandwiched between them was Rashod Bateman (3-2-15), who had the most notable play when Lamar Jackson threw an INT his direction, on a pretty simple comeback/stop route on the sideline where there was a clear miscommunication on the timing and positioning coming out of the break. Jackson had his hands up on one replay sort of like “What was that?” The vibes around Bateman continue to be not great, and it’s possible his play isn’t allowing them to ramp up his usage. Flowers is the only WR you can even consider trusting in the short term here, which is just to say that not much changed here in Week 6.
Ryan Tannehill was knocked from this game early, though he was around for three quarters of bad football. DeAndre Hopkins (5-1-20) could only muster the one catch, Nick Westbrook-Ikhine (4-3-25) and Chigoziem Okonkwo (4-2-18) weren’t any better, and it wasn’t until Tyjae Spears (4-15, 1-1-48) got loose for a late big play that was basically all YAC — and came on a pass from backup Malik Willis — that the Titans did anything notable in the passing game. Spears continues to be super interesting, but the Titans’ offense continues to be super uninteresting, which is a conundrum.
Derrick Henry (12-97-1, 2-2-16) also got loose for a 63-yard run earlier in the game, but he did little else the rest of the day, totaling 34 yards on his other 11 rushes.
Signal: Gus Edwards — 61% snaps, 49% routes, all 3 RB green zone touches (strong role despite not getting much going in a tough matchup)
Noise: Titans — 46 plays, 21 passes (there’s Signal here, too, because the Titans are obviously not a team that runs a lot of plays or throws a lot of passes, but this is extreme low volume)
Commanders 24, Falcons 16
Key Stat: Terry McLaurin — 1.13 WOPR (second highest in Week 6, through Sunday)