Week 1 is here, which means every league chat is about to fill with the same question: which quarterback do I start? Yahoo dropped its Week 1 quarterback rankings this week, and per usual, the top of the board writes itself. Josh Allen, Lamar Jackson, Patrick Mahomes, Jalen Hurts. If you drafted one of those guys, you're not reading this paragraph anyway, you're setting your lineup and moving on with your Tuesday.

The interesting stuff lives in the middle, where the rankings get fuzzy and the decisions actually cost you games. That's where I live too, so let's dig in.

The classic Week 1 trap is overreacting to matchups you haven't seen yet. A QB facing a "soft" pass defense from last year isn't necessarily facing a soft defense this year, because defensive personnel turns over more than any position group in football. New coordinators, new corners, rookies stepping into starting roles. The version of a defense you're evaluating in your head is September of last season, and that team doesn't exist anymore. I'd weight the talent of my quarterback over the perceived matchup every single time in Week 1, and only start letting matchup data pull me around from Week 3 onward, when we have actual 2026 film.

The other trap: benching your starter for a shiny streaming name. If you drafted a guy in the top 10 at the position, start him. Week 1 is the single worst week of the season to get cute, because you have zero data on your own guy and zero data on the streamer. Deb yelled at me about this exact thing two years ago when I benched a borderline QB1 for a matchup play and lost by four points. She still brings it up. She was right, which is annoying.

Where streaming does make sense in Week 1 is the bye-week-in-advance crowd, the managers rostering a QB2 they'll need in a few weeks anyway. Grab your guy now before the waiver wire turns into a stampede. The best Week 1 streamers are usually the ones nobody's scouting yet.

One process note on how I use rankings like Yahoo's: they're a sanity check, not a lineup setter. I've said this before about model outputs and it applies to human rankings too. If Yahoo has a guy at QB14 and I have him at QB8, that gap makes me go back and check my work. Maybe I know something about target share they don't. Maybe I'm just being a homer about a guy I drafted in three leagues. Rankings are good for catching your blind spots, not for outsourcing your decisions.

Ray would tell you to just watch the games and trust your eyeballs, and sometimes the old man has a point. But in Week 1 nobody has eyeballs on anything yet. That's exactly why the boring play is the right play: start your drafted starter, don't chase last year's matchup data, and save your risk budget for the weeks when you actually have information.

I'm logging every start-sit call I make this season for a midseason receipts column, including the ugly ones, so if you follow a recommendation from me this week, know that it's going on the board either way. Who are you sweating at QB this week, the guy you drafted or the guy you didn't? Drop it in the comments. I want to know who to say I told you so to.