ESPN's FPI put out its top 15 college football championship contenders going into Week 2, with Ohio State, Texas, Georgia and Michigan all in that conversation. My honest reaction: it's a beautiful photograph of a room that is about to get rearranged by furniture movers.
I don't have the full ordering in front of me yet, and I'd argue it matters less than the shouting suggests. A rating system in the second week of September is mostly preseason belief with one weekend of evidence sprinkled on top. That isn't a knock on the math. It's just what the math has to work with.
Who has the best chance to win the college football championship?
Every projection built this year has Ohio State and Texas living near the top, which is exactly why their No. 1 versus No. 3 non-conference meeting headlines Week 2, and exactly why coaches keep repeating that these games are the only way to define yourself as elite anymore. Georgia stays in the picture the way Georgia always does, with a Western Kentucky tune-up at 12:45 ET Saturday that will tell a model almost nothing it didn't already assume.
checked the almanac: september 10, 2005, texas walked into columbus and beat ohio state 25 to 22 with vince young running the show. ohio state went to austin the next year and won 24 to 7. twenty one years to the day, same two programs, same top-of-the-poll billing. college football keeps its receipts.
Why Michigan's quarterback question outweighs the model
Michigan is the one that makes me squint. The reporting out of Ann Arbor says the quarterback problems run deep, plural, not a single bruise, and one analyst on the network framed this as the biggest game of Bryce Underwood's life. Oklahoma comes to the Big House at noon ET Saturday with John Mateer, whose offseason in the weight room became its own genre of content.
Here's the thing a projection system cannot do. It cannot feel a quarterback room. It sees returning production, recruiting grades, last year's efficiency, and it hands you a tidy percentage. It does not sit in a Tuesday practice and watch a coaching staff decide who takes the first snap. Models don't get nervous.
If Michigan's passing game looks unsettled against a real defense, the list gets rewritten from the middle out, not the top.
What Week 1 already taught the ratings
LSU is the cleanest example of how these systems actually learn. Beat Clemson in Week 1 and you climb SP+, because one result against a genuine opponent moves a rating more than three results against overmatched ones. That's not favoritism. That's information.
Which is the quiet problem with grading Week 2. Look at the slate: East Tennessee State at North Carolina, UT Martin at West Virginia, Howard at Indiana, Holy Cross at Miami of Ohio. Nobody learns anything from those. The whole weekend's evidence gets carried by two or three games, and the rest is bookkeeping.
One more wrinkle worth filing: the NCAA has implemented new standardized timing protocols this season. Fewer clock stoppages means fewer snaps, and fewer snaps means fewer chances for the better team to prove it's the better team. Every ratings system in the sport is built on per-play efficiency. Shrink the sample inside each game and you get noisier weekly movement, all year. That's a tell about why these lists will jump around more than usual.
And tucked underneath it all, Morgan Scalley is coaching his first year at Utah, a former Ute defensive back who now has to drag his alma mater into the playoff conversation. Those stories don't show up in a top 15 in September. They tend to show up in November.
What I'm watching next
Three things. Whether Michigan names and trusts a quarterback by Saturday noon. What Ohio State and Texas actually look like when a top-three team punches back, because that single result will move the projections more than the rest of the weekend combined. And how far LSU keeps climbing if the Clemson win turns out to be the real thing rather than a Week 1 mirage.
The list is a starting point, not a verdict. Saturday does the actual grading.