Stealing Signals Week 4 Introduction

A one-off post before I dive into the numbers

I get pretty passionate sometimes about this stuff, and it’s in part because I’ve made the mistakes. I started playing fantasy football in 1999, but I wasn’t doing anything all that special until long after that. It really wasn’t until I started writing this column at RotoViz in 2017 that I started to figure it out. When I started making specific weekly predictions about every team, every week, and feeling like the data would tell me everything, I started to find out I was frequently going to be wrong, almost no matter what, and more than I would’ve have guessed I was based on my results prior to that.

I started writing this introduction to Stealing Signals this morning and it got so long I’m going to make it its own post. I’m not sure why I’m writing this instead of what I’m supposed to be writing, but that’s more or less why all of you subscribe anyway. I’ve learned my best work comes from leaning into this idea that if I want to write about something, I should just write about it.

Sometimes in football there are no concrete answers. It’s a week-to-week league, they say, where teams can look very different each Sunday. Stuff evolves, gameplans evolve. And I think there’s so much information overload on Sundays that observers have a tendency to cope a little bit by blocking out how wrong they were on some of the negative stuff and focus more on the positive. There are several biases that play into that, but hindsight bias and confirmation bias are the strongest, I’d guess. It’s really easy to think after the fact that one situation that went the way you thought was always clear, and another situation that didn’t was always more muddy, and you should have recognized that.

But what I quickly learned when I started writing this column is no matter how hard I tried, no matter how much effort I put in, I wasn’t going to find every answer. I don’t want to get that confused with extra work isn’t helpful — it obviously is, which is why I still write the column.

But the thing this column did when I started writing it was it more or less forced me to boil down my thoughts to takes about every team, every week. Longtime readers will know the shrug emoji has made its way into my pieces in the past; I would like to give that sort of take far more than I do.

And what I learned when I started really trying to lock in my conviction and really do the best analysis I could do rather than just throwing my hands up — there’s a lot more you get wrong than you think. This is a big reason why I get pumped when I’m right, and for the right reasons, like how I described D.J. Moore’s upside scenario here. In an industry where victory lapping is common — and those victory laps are very often a result of getting a call right but for opaque or sometimes straight up wrong reasons, because things really do change so much, and because this game really is humbling, and because no that analyst who acts like he’s always right is very much not just that sharp, he’s just good at marketing — there’s obvious reason to celebrate sometimes when you go against market and hit something right on the screws and knock it out of the park.

But the point here is there’s a lot more you get wrong than you’d expect, and your stupid, stupid brain convinces you that’s not true. It convinces you that what you got right was always predictable, or if you were way off then in those cases it was just random. You get overconfident quick, or if you’re on a cold streak you can lose that confidence just as quick, feeling like you have no grasp on anything. If you actually care about this stuff — if you’re playing DFS and betting sports and trying hard to win at all these different ways you can analyze football — the single best thing you can remember is nothing is what it seems. Chaos, remember? Nexus events, and all those other ways I tried to describe it? NFL seasons are chaos.

But I still got the questions about D.J. Moore’s touchdown-scoring ability, like we know something. And I don’t think those answers are all solved; I just knew and still think we didn’t know anything, and were overemphasizing what we’d seen. And I’m seeing in my new subscriber Discord — and shoutout to you guys for the awesome conversations there, it’s fantastic — that there’s the occasional post or two that I would just wish were a little less certain about everything.

There’s this phenomenon called the Dunning-Kruger Effect I reference from time to time. It looks like this:

It shows how people often are initially overconfident in something, then find troubles and have to adjust before developing more competence in that thing. And once they do start to find all those more complex elements and feel secure about them, only then, after being humbled, does their confidence again grow.

I only just realized this week that the fantasy football Dunning-Kruger chart doesn’t look like this, in a very important way. The sharpest people I know in this space — the people whose opinions I trust the absolute most — understand how humbling this is.

The key difference looks something like this:

As competency reaches its highest form, confidence never gets back to the levels it was at the Peak of Mount Stupid. And what that means is in this industry with lots of turnover, the loudest voices at any given time are typically newer folks, not yet humbled. I sound like such an asshole writing that, but I do believe it’s true, and I do believe it’s a valuable point to make. In my time writing about this, I’ve watched people come and go, and I’ve seen confidence shattered, and I’ve seen new analysts with very confident takes absolutely nail stuff right out of the gate and look like geniuses, even for extended periods of time, but eventually everyone gets humbled. That’s how this works.

I wanted to write this because I frequently get asked questions that I very much do not know the answers to, and no amount of thinking hard or analyzing is going to get me any more confident than maybe 60/40 in a lot of those discussions. And because in this early part of the season, I see that confidence out there. And again, I am being such a boomer, but I legitimately do want to save some people from themselves. It’s so tough and as I started at the beginning, I’ve been there, getting a little too out over my skis, and the first time you crash in a big way with attention on your takes it can really be tough. That Valley of Despair can be rough. And sometimes that extends. Because there’s so much variance in this, sometimes you can just run cold for a bit, and man, if that comes at the wrong time, and stuff builds up, it can start to feel like you’re wrong about literally everything. And then people start changing up their whole processes, and chasing dragons, chasing that high of when everything felt so easy. Or else people just turn away and they quit on the whole thing. And that just sucks. It’s an issue with expectation.

So anyway, I just wanted to address that expectation gap, as I see it. We keep getting more and more data, but football isn’t solvable. There are way too many layers here, from the interconnectedness of 22 players on the field to the schemes that dictate how they move and what they do, to the balance between player skill and opportunity, to whether player skill atrophies with injury in violent sport, to so much more. Some stuff is actually pretty clear — some teams use the same schemes each week, and their player roles are pretty secure, and some of those roles might even stay consistent year over year — but man, that’s like 20% of the league. The overwhelming majority is shifting, somehow, and shifting always. And when you understand that — when you truly accept that you shouldn’t be overconfident about anything — that’s when you can figure out how to play the market, and that’s when you can find an edge. You still have to understand how to find that edge, and identifying good players and all the things I talk about in Stealing Signals still very much matter, but it’s always conditional. You act when the timing is right; when there is too much certainty one way and you can make that bet at better odds than whatever your confidence in it is.

That’s all this game is — seasonal, DFS, dynasty, betting on football lines, betting on props. And believe me, when I write my Signals and Noise each week, I’m very open to being wrong. The most important thing I always try to remember — especially now, when I feel great about D.J. Moore and about several other calls — is how very humbling this can all be.