Every process has holes

Some initial thoughts on the 2025 season

I have so much to write. I’m pretty overwhelmed, but not in a negative way, because there’s time. In fact, I’m thrilled about it.

From a practical standpoint, I’m aware my approach in 2025 to wear some personal stuff on my sleeve is going to cost me some subscribers longer-term, but my mindset with that stuff is always that if I just do more good work, I’ll attract more subscribers to replace those who left. That’s how it’s always worked; I have confidence in my ability to bring a unique voice to this space.

The issue is you can’t just manufacture the kind of good work I’m talking about. You have to have something to say. That’s where the thrill in being overwhelmed lies for me. I’ve learned for these kinds of things, to jot down what I can, and then I can read it back later and try to get back to the headspace of turning these unstructured notes into a piece. That works most of the time, although sometimes the half-baked ideas are just half baked, so I leave them.

I have a simple Google doc for this, and a month or so ago I pushed some stuff down that doc to make room for my initial thoughts on what will eventually be my “biggest takeaways from 2025” piece, because it’s become a tradition around here for me to list my three or sometimes five biggest thoughts about the season, and the sport. I always talk about that quote from Siegele that’s technically under my byline but he wrote as a concluding sentence to a piece I’d put together along this path, early in my career. It remains such a banger, and such a concise and accurate way to summarize a massive part of what I was saying in a post over thousands of words.

“Drafters almost always create themes out of last year’s most compelling stories — even when those stories do not reflect the most important trends — and they struggle to identify future shifts in style or opportunity in any meaningful way.”

Those posts are coming. But I mention them to say that I’ve written so much other stuff over the few weeks since I started the brainstorming for that post that those bullet points are now 11 pages down in this Google doc I have. In fairness, about half of that is notes from the games in Weeks 17 and 18, which I did in my normal style to help me figure out which games I’d even worked through (I’ve only seen about half), and then there is some to-do stuff from a business standpoint and other stuff. But I mean there are a few pages of content and ideas already written. There are like four solid pieces in there, and I’ll do my Field Tippers pieces, and I might try to do something similar for RBs (and QBs) this year — there’s so much to write about.

So I’m excited. I’m very competitive, and my mindset is something like, “I don’t typically generate a lot of subs in January, but I’m going to this year.” And there’s a part of me that views the whole thing as a necessary cycling. I like attracting new readers, and while I’m never trying to chase people away — and I’m certain some of who I’ll lose after 2025 (I lose people every year, in some capacity, for a variety of reasons, there’s no such thing as 100% retention, so this isn’t some big deal) will be subscribers I’m bummed to lose and wish I could’ve kept around — there are those who won’t be back for 2026 that wanted me to do fantasy content they were more used to, and weren’t a great fit. The way I approach this is always going to be unique — mostly because of that comment that I think the content that matters is when you have something valuable to say, and the content most people associate with fantasy football is far too formulaic to fit that — so cycling out some of the subscribers who aren’t looking for that unique stuff, and want me to be more like other fantasy football content producers, isn’t something to be upset about. Especially if I can also cycle in potential longer-term subs who share you are readers for life, and those kinds of nice comments many of you have shared, because you really value the way I approach it and find it to be indispensable.

I have so much to write this winter for another reason. A lot of what I’ll write will touch on things I touched on in the intros, but maybe didn’t fully form, or I just want to touch on again.

One thing I wrote about at some point was this idea that evolutionary processes don’t really have an ending point. We like to believe the moment in time where we sit is the thing everything was building toward, but that’s never it. I’ve written so much about macro stuff with this sport since, well, I started writing in 2015; I recently looked up and referenced that Infinite Jest quote from David Foster Wallace about “the aleatory flutter of uncontrolled, metastatic growth,” which came from my big 2017 RotoViz post on macro trends, where Siegele had the above quote in the conclusion.

But more recently, I’ve written so much about macro stuff since the defensive focus shift in 2021 and 2022, and the wait for an offensive counterpunch going into 2023 and then 2024, and then now we just finished 2025 and it was a season where so much happened in this evolutionary span. The sport is evolving before our eyes, constantly. We’ve been in this period of unrest, and instead of eras where the sport changes in smaller ways slowly over 5 or 10 years, we’ve been in a period of fast innovation where there have been multiple different shifts in just the past three or four years, sparked by teams approaching things more optimally overall, and breaking rules they once held sacred.

Pardon my language, but I’m going to write so fucking much about all of that, and about what it means for fantasy, and how I think a ton of what we’ve done needs to shift, and which direction, and which other tentpoles need to stay firm even if the results this year tried to tear them down, and why. I’ve stepped back from some of my podcasting and other stuff, and this is just an aside but I have noticeably less influence in my industry going into 2026 than I did even in like 2024, and I absolutely love that. There’s a lot of stuff being discussed that I think misses the mark, and there was a stretch in my career where enough people read everything I wrote that I felt like I would never again have that “outsider” ability to just react to that stuff, which sucked because that’s what I do best, and where I need to be positioned. But I think we’ve kind of circled back that direction some, and again, I’m thrilled.

One of the things that happened last year is for the first time in a really long time, I didn’t go into the summer with one big theme, or one big piece I’d kind of chewed on for months and wanted to get together. It scared me a bit — I wondered if I’d kind of covered everything I wanted to cover in this space, and was going to be just regurgitating takes and overly self-referential going forward, which is not where I want to be. Again, I’m not the analyst that can teach an audience the basics over and over again, from scratch, every year. I can’t make the same points repeatedly, just updated for the 2026 landscape. It burns me out.

But as I wrote in that November post where I quoted Infinite Jest — which I just have to admit, sigh, that quote came from the first third or so of that book, because I’m one of those people who started but never finished that famously difficult read, which just adds so much embarrassment to the pretentiousness of quoting David Foster Wallace — 2025 was what I called in the title, a “postmodern, transition” year.

Again, the thing about evolution is there’s no stopping point. We’re in it. And while you’re in it, the foundational stuff is always tested as the landscape changes. Last year maybe felt too similar to the couple years before, and the big counterpunches maybe took a year or two longer than I anticipated. But offenses started doing Cool Shit this year, and the whole sport just keeps changing. It feels like I have multiple offseasons’ worth of the types of posts I said I didn’t really have in the bank last year; 2025 was just so refreshing.

---

Today’s theme is more reflective. I talked about my process feeling like more of an outlier these days. I think that’s the result of the shift I consciously undertook a few years ago in how I watch the games, and then doing less podcasting overall, although it actually wasn’t ever my intention to not have a regular podcast between the end of Stealing Signals Tuesdays and the start of the following week.

I left Ship Chasing last year, and lost what had been a Thursday night pod, but I’d always recorded Stealing Bananas on Thursday mornings. For a variety of reasons, Bananas became a Sunday night only show this year and I went from multiple late-week recordings in 2025 to zero in 2026. That’s something I’ll rectify in the future; I didn’t need to overdo it this year, for reasons I’ve given, but it wasn’t my preferred way of engaging with the season.

But I think that, as well as the continued destruction of social media, led to this isolationist feel, where I wasn’t seeing as many other opinions. And then what would happen is I’d eventually see something from someone I don’t really consider a hot take artist, who is more plugged in and engaged and wouldn’t talk about stuff without feeling confident in it, and they’d share something as if it was a known known, and it would just shock me. I’d feel not like that person was wrong, but that how casually they were giving that perspective meant that a lot of people were thinking about something very differently than me.

It doesn’t mean I was right, either. There’s this fascinating thing with football where the way you engage with a season builds. Take my Tank Bigsby position from 2024. I felt early in the year that Bigsby’s peripherals were juiced by a couple big plays that basically any back could’ve hit, and then the game around Week 4 or whatever where he had double-digit Missed Tackles Forced against the Colts. And from there, nothing Bigsby did all year refuted that position. I spent most of last offseason out on Bigsby because I had a position through 2024 that was different than others early, where I was skeptical and needed to see more, and then I felt we very much did not, and so I felt my priors — that he was a decent but unspectacular prospect who had a horrendous rookie year in 2023 and probably wasn’t a difference-maker — were correct, and didn’t need to be adjusted.

I may still be wrong on Bigsby, who has looked OK at times in Philadelphia, but the point is I had a very different opinion on him than most of the industry, and I think it stemmed from this idea that a lot of people who were looking heavily at the data, and every week they were updating the seasonal efficiency leaderboards, were seeing Bigsby in a favorable light week after week, and told themselves a story he was just frankly much better than Travis Etienne in 2024. Etienne looked like a guy who was fading, as far as the peripherals had it, despite the evidence we had he was clearly playing through injury, and our knowledge that he wasn’t all that old, and didn’t have all that many career touches for a former first-round pick with a much better collegiate profile and two 1,400-yard seasons at the NFL level. I’m never all that bullish on veteran RBs, and this is one of the things I do want to address this offseason, because Etienne was a guy I basically only wrote about favorably, and yet I still missed on him to a degree because of the presence of Bhayshul Tuten (which, in fairness, a third RB did complicate this analysis, and open up multiple paths to being wrong on Etienne).

(I say “to a degree” rather than in full because I did draft him in a couple leagues, including one of my two long-running home leagues where friends tend to steal some of my other targets, but that’s sort of the overall issue is I didn’t get much because I always liked an alternative, and you guys likely were in the same boat if going off my ranks, so that’s one in hindsight I do feel like I should have had ranked more aggressively just based on what I thought about the player and situation. It’s the kind of thing I’ll address more going forward, as well.)

Anyway, that was all a tangent on the point about Bigsby from 2024, and I’ll continue it real quick to say the best comp for 2025 is Woody Marks, who similarly had some early-season big plays that every other back could’ve made, and a big performance against an overmatched opponent in Week 4, and then was pretty poor the rest of the year, routinely failing to hit his weekly projection despite probably the weakest backup RB competition in the NFL, and the Texans getting hot and winning a ton of games, always putting him in theoretically great game scripts.

Some of the issue there is the Texans drafted a weird back for what they need, since Marks’ skillset is as a pass-catcher but this is a run-and-defense team with a QB in C.J. Stroud who pushes vertically and doesn’t really get to his RBs for checkdowns all that often, so Marks was asked to be more of an early-down grinder, which is the part of his prospect profile that wasn’t just NFL replacement level bad but was clearly and blatantly not even at a draftable prospect level. He’s not a good runner! He played on multiple college teams and got decent work across five college seasons, but never posted good MTF, Yards After Contact, Breakaway Rush, or even rush attempt numbers, as no team was ever all that thrilled about giving him a ton of running volume. There’s basically nothing about his 2025 that refutes any of that, and similar to Bigsby last offseason, we go into this one where I’ve heard a lot about Marks from a number of sources that just doesn’t fit with anything I came to believe about him for the 2025 season.

There’s a lot I missed, too, so this isn’t some criticism. I’m all over the place with this post, but the theme was going to be more about my own process. Years ago, when I started writing Stealing Signals, my focus was on the data. I watched the games on Sundays, but I watched them all at once. I got a feel for things that were happening in every game, but at the risk of sounding pretentious again, there’s nothing like watching a full game from start to finish.

To that point, in Week 17 I joined the guys at Ship Chasing for their ShipCast, which was such a fun thing I did the past couple years and I missed this year, watching live football and talking to people I enjoy chatting with and all of that. I got an idea of all the games, but honestly, I was pretty overwhelmed. I’ve made a point to go back through some of those Week 17 games, and I watched a ton of football yesterday even though it was just Week 18.

But because of this newer process where I normally detach on Sundays, and make a point to watch every game all year from start to finish — and really understand how the game went and why — I’ve found in more recent years that Stealing Signals is less in tune with the data. Early in my career at RotoViz, I was more of a data nerd, and there were film vs. analytics debates, and I was always so annoyed at the film people who acted like they knew so much more about football because they watched the all-22 on Mondays, when their fantasy football records in public leagues they participated in suggested an inability to take that information they were sure everyone else was missing and apply it to the actual hobby in a successful way. These days, I have more in common with the film grinders of that era than some of today’s data slicers.

Back then, there was a lot I’d miss. I could tell you every stat about every player and team in the league, basically on a weekly level, and yet my process had holes. I’d miss on stuff like Bigsby over Etienne, and it was hard for me to understand why. And for what it’s worth, the why on that one really does come down to just ignoring what the data said, on that one. Why? Just because when you really watched, you understood the data came together in a misleading way.

I sometimes use this phrase “it’s small samples all the way down,” and this stuff is why I prefer the simple data. I’ve recently been annoyed at EPA, which has come to dominate the “real football” and even some fantasy football discussion. The thing about EPA is it’s a good stat that’s directionally accurate that nonetheless catches a ton of noise. And then people who act like they are the only ones who know the real truth about the sport of football get very mad when you look at a full season sample of EPA and there are rankings that put two players back to back with a slight gap, and you say, “This is within the error bar.” Variance should be the expectation with EPA. It only looks at the result of the play. For a QB, getting away with dropped INTs, getting a fortunate deflection or a great play by a WR — or having a WR drop a pass — these things are all missed by EPA. It’s stats like EPA that are why places like PFF were formed whatever it was, 20 years ago.

When you look at EPA stats, you should think of it like the real value is somewhere within an error bar that stretches plus or minus whatever standard deviation or figure makes sense, assuming you’re not going to actually chart every single result, and whatever you do, you absolutely without question shouldn’t be overly precise about the specific end number you have. It’s directionally accurate but not specifically accurate.

This doesn’t even get at the fact that EPA on a per-play basis ignores the actual play-calling, and how I’ve said “volume is efficiency,” which I believe I stole from Adam Harstad, and is a fine shorthand but isn’t really the point so much as something like, “What a lot of volume tells is the same as what great efficiency tells us, which is that the player is really good.” I remember back in the day, with the old Titans offenses, how EPA-based information was being hammered to say Derrick Henry was not as valuable as Ryan Tannehill, and I did some digging and there are probably tweet threads out there, but my argument was Henry got hit for negative EPA plays in situations like a second-and-18 where they’d call a run, and he’d gain 7 yards and do what he could in that instance to get them into a third-and-manageable, but it was still a negative EPA play because you’re still at third-and-11 and all EPA looks at is the before and after snapshot. And my argument was that Tannehill was only throwing in spots where he had the potential to gain a lot of EPA but not really lose it, like third-and-longs, or else he was throwing off the heavy rushing gameplan where play-action and the realities of nonobvious passing downs gave him a real benefit.

The lesson there was that EPA per play was not factoring in the volume is efficiency concept that the Titans were giving the ball to Derrick Henry every single time because they couldn’t actually try to throw out of second-and-18 and go get a first down. The Chiefs will always put it on Patrick Mahomes on second-and-18 or any other down, and that’s shown up across his career in terms of pass rate but has also misled people into believing he isn’t as good as other QBs at times, because of per-play metrics that don’t understand how the sport is actually played.

This year’s Derrick Henry and Ryan Tannehill? Obviously the Seahawks, with their massive run rate despite the lack of success for much of the year, and then people kidding themselves into believing Sam Darnold was a top-five QB for a large stretch of the regular season since his EPA per play data was literally league-leading for a good portion. But beyond the initial stuff I said about variance and good fortune and those things — and beyond the obvious of Jaxon Smith-Njigba creating more separation than any WR has any right to, on every damn route concept — there was this point about Darnold not actually being relied on to go get the tough yards through the air in the difficult passing downs, because the Seahawks couldn’t deal with a back-breaking INT on third-and-long to give up field position their defense could hold if they just ran the ball, took the negative EPA rush, punted, and then got a stop defensively and tried again.

That’s the way they play football, and it’s worked. It’s not unique to them, or the Titans back in the day, and it doesn’t mean that the average pass still isn’t a better proposition, and teams run more than they should in general, which is why so many more runs have negative EPA than passes. But it is necessary context of the stat that the way the game is called influences the per-play efficiency. If your QB is only dropping back on nonobvious pass downs, chances are his EPA per play is going to be inflated relative to someone like Mahomes where every dropback is an obvious pass down and a clear passing defense look.

Anyway, EPA is a viable shorthand, and I’m not trying to kill it. It’s not gospel, and a new generation of football fans needs to learn that every few years, but as long as you understand it as a directional representation of what’s happened rather than an exact one where ranked lists are definitive (which is the exact wrong way to use a stat where our expectation is the end results are influenced by variance), you’re fine. It’s a great tool.

The point of this post is supposed to be that that’s the case with every football stat. We don’t realize it, in real time, and it’s so easy when one process doesn’t work that we look for another that got that one thing right. But I promise you if you chase that other process, you’re going to find other holes.

I had holes this year on the data side, I think. I wanted to do better with understanding some of the advanced usage stuff, like the rate of horizontal breaking routes which seems like a legitimately fun and exciting advancement in charting data for WRs and TEs these days, given how much more lucrative those specific routes tend to be, statistically (the use would be understanding who runs those routes in which offenses, and to essentially expect them to have higher TPRRs as a result).

But in the past, I had holes on the film side. I do think I bridge the two well, and I still look at a ton of data. I just find that my Sundays are way longer these days, and thus my Mondays are trickier, and I don’t find it as easy to approach Stealing Signals with the wide-eyed enthusiasm of picking through all the data from every team, as I write. Instead, I’m fixated on my diligent notes from watching the games, helping to explain the specific box score more than understanding the complexities of some of the more advanced numbers in every case (I still try to do that, but again, holes). I write so many games every week and the issue was realizing at times I didn’t even check something simple — say, a key WR’s air yards profile — as I worked through one game, because I didn’t find it all that relevant to the story of that game. And just knowing that in the past, my process would’ve never let that specific piece of information, that I did find relevant, to slip through the cracks.

There’s is inherently not enough time to analyze this sport every week. I really don’t know how one could actually watch every snap from all the games, every week, and also go through all the data, every week, and do the work of meaningfully interpreting and analyzing it, every week, and not have holes. There will be people who will tell you that’s their process, and probably there are some where that’s true, for a time. I’d be pretty skeptical of anyone who claims they can keep that up, every week, for four months. And then if they can for a season, that they won’t break down over the years, and probably just after two or three, because I promise you that’s a gauntlet to do week in and week out for the length of a season.

I do like where I’m at now, and the point of this post in some ways was to talk through that evolution and explain I do feel my current approach is the best thing I can provide. There’s so much context that’s lost when you don’t watch a game in full; for whatever reason, I’m thinking back to the Week 17 Sunday Night Football game, which I didn’t see live because none of my results hinged on it, but I followed it and saw Luther Burden had a big day and the 49ers won a shootout and Brock Purdy put up some numbers. And I had a certain perception of that game basically all week, but when I finally got a chance to watch it a few days ago, I think on Saturday, there was just such a different feel to how those stats were accumulated than what I thought. I don’t really know how to articulate that any differently, other than to say it’s a different vibe, because I’d stop short of saying the additional context is always better; it can at times lead you the wrong way.

I wrote a few years ago about how so much of the data I used to look at in the formative years of Stealing Signals is now ubiquitous across the fantasy landscape, and one of the things that’s happened further since then is certain processes being held up as the right way to break things down, in a way that is too confident for my taste. People talk about player usage data as if it tells certain stories that are sometimes just patterns in the numbers, and for some of the takes I see, the response really does come down to that old film vs. analytics era retort of, “Do you even watch the games?” People are sometimes so confident of things that are not just probably wrong because we have a difference of opinion, but are demonstrably incorrect if you literally just watched the sport more.

Anyway, I’m mostly just musing about this, as it’s where I wanted to start my reflections on the season. I don’t think there was a ton of meat in this post, and there will be a lot more specifics in the next ones I write, which will ironically be data-intensive. Hopefully that’s not lost in my references to the old film vs. analytics debates — I’m far too obsessed with data, too much of a strategist, and way too competitive generally to lose the curiosity about what actually wins. It’s way more about the data becoming ubiquitous, and the processes by which people play fantasy now being so different, and leaning into where there are edges. (The truth about the various versions of the film vs. analytics debates has always been you need at least some degree of both to properly contextualize, and I do hope I have the ability to analyze both sides of it well enough to find good answers more often than not.)

As I get more into my process reflections, both thinking about where the sport is headed and diving into the data that was accumulated, hopefully it’ll make sense why I started here. It was great to hear that a lot of you finished well this year — for anyone curious, we did take down the $5k, I won one of my dynasty leagues, Shawn and I had a team finish right around 50th overall in the Main Event, I had the earlier FFPC division titles, and kind of a fun one was I won one of my two 18-person Fantasy Cares Eliminator leagues — but I know a lot of you came up short, as did I in a lot of leagues I wanted to win.

That doesn’t mean I’ll reinvent the wheel. But there are a lot of things to consider. If you’re on social media or plugged in at all, you’ve likely seen a lot of people celebrating successes, and if you didn’t have them, that can be tough. The urge is there to try to figure out what those people know. But don’t miss the forest for the trees, and don’t be sold by snake oil. Every process has holes. The goal is to find success despite them.