Week 1 of the 2026 season is complete, and the College Football Playoff predictor has already turned one weekend of football into a season-long verdict. That is the job. It is also the reason you should hold every one of these numbers loosely for another three Saturdays.

The machines came out in a bunch this morning. ESPN's Allstate Playoff Predictor posted its Top 25 by playoff probability, the FPI top 25 landed with LSU and Texas flagged as the biggest Week 1 risers, SP+ refreshed all 138 FBS teams, and the network's human panel voted seven different SEC schools into its playoff projections. I do not have the full 25-deep probability list in front of me yet, so I am not going to pretend to rank it. What I can tell you is what the inputs are, and where the models are quietly fighting.

What the College Football Playoff predictor liked after Week 1

LSU is the headliner. The Tigers went out and beat Clemson in the opener, and the FPI responded by bumping their title chances, which is exactly what a computer is supposed to do when you take a marquee non-conference scalp in September. Texas got a similar nudge. Miami, Ohio State and Indiana all made statements strong enough that the poll conversation shifted around them, while Oregon and Michigan are the two names being discussed as candidates to slide.

Seven SEC teams drawing playoff votes from the panel is the number I keep circling. Seven. In September. That is not a prediction so much as a confession that nobody in that room wants to be the one who left the eventual champion off the list. Panels hedge. Models do not, which is why the two are worth reading side by side rather than one instead of the other.

Why Iowa is the most interesting name on the list

Iowa beat Northern Illinois 40-0 and the FPI essentially shrugged. That is the kind of result that makes fans throw a remote, and honestly the frustration is fair on its face: 40-0 is 40-0. But it is also the most defensible thing the model did all weekend. Predictive systems discount blowouts of overmatched opponents on purpose, because the margin tells you less about the winner than it feels like it should. The Hawkeyes get to argue with it in the Cy-Hawk rivalry, against an opponent the computer actually respects. That is where a shutout gets converted into a ranking.

checked the almanac: the ap poll has been ranking college football teams since 1936, which means voters have been wrong about the first weekend of the season for ninety years running. the machines just get to be wrong faster now.

The LSU asterisk nobody can model

Here is the part no predictor has a column for. While LSU's on-field arrow points up, the SEC filed an amendment to its federal lawsuit against LSU, Lane Kiffin and school officials, seeking the authority to expel LSU from the conference. I am not going to speculate on where that lands or how fast, because the reporting does not support a guess and neither do I. But a playoff probability built on schedule strength and conference path is a probability built on assumptions about what conference a team is in. That is not a knock on the model. It is a reminder that a model can only chew what you feed it.

What I'm watching next

Three things. The Week 2 AP poll, which will show whether human voters follow the computers on LSU and Texas or dig in on Oregon and Michigan. Oklahoma at Michigan on Saturday at noon ET, which is the single biggest resume swing available this week and the fastest way for a Week 1 doubter to look smart or foolish. And Oregon at Oklahoma State, also Saturday, where a Ducks team the projections are already nervous about gets a road trip to prove the nerves wrong.

One weekend of data, four different systems, and they already cannot agree on Iowa. That is college football working exactly as designed.