ESPN's FPI top 25 rankings came out after Week 2 with Ohio State sitting at No. 1, which would be an ordinary sentence except the Buckeyes just blew a 20-point lead to Texas. That is not a bug in the machine. That is the machine telling you exactly what it was built to tell you, and everybody hearing something else.
The internet did what the internet does. The word "controversial" got attached to the poll before most people finished their coffee. I understand the reflex. You watch a team lead by three scores, watch it come apart, and then see that team at the top of a list the next morning, and your gut says somebody is not watching the same games you are.
Your gut is wrong here. Not about the collapse. About what the list is for.
Why is Ohio State No. 1 in the ESPN FPI top 25 after losing to Texas?
Because FPI is a predictive metric, not a resume. A human poll answers the question "who has earned the most so far." A predictive index answers a different one entirely: "if these teams played next weekend, who wins." Those questions can produce wildly different answers in September, and the distance between them is widest in exactly this scenario, a team that was demonstrably better for three quarters and then was not.
A 20-point lead is not nothing to a model. It is evidence. It says that for a long stretch of football, one team moved the ball and stopped the other one at a rate that looks like a national title contender. The collapse is also evidence, and it counts, but it counts once. It does not erase what came before it. The scoreboard is binary. The play-by-play is not.
There is also the calendar. Early in a season, these formulas are still carrying preseason expectation, because two games is a sample that would embarrass a coin flip. That weight fades as the season fills in. It is heaviest right now.
FPI, SP+ and the human polls are answering different questions
Same morning, the updated SP+ numbers landed for all 138 FBS teams, and they come from the same family of thinking: grade the efficiency, strip out the noise, project forward. Both of those live in a different universe than the AP ballot, and they are supposed to. The problem is that we publish them in the same shape, a numbered list from 1 to 25, so readers treat them like competing verdicts on the same trial.
They are not. One is a jury. One is a weather forecast.
checked the almanac on this one: in 2014 Ohio State lost at home to Virginia Tech in September, got buried by every human poll in the country, and finished the year as the first College Football Playoff champion. the eye test and the résumé were both screaming the same thing in week two. both were wrong.
What this means for Ohio State and Texas the rest of the way
For the Buckeyes, the actual stake is not a September ranking anybody will remember. It is whether the fourth quarter was a fluke or a fingerprint. Blown leads have a way of following a roster around, less as a curse than as a thing the team starts thinking about while it is happening. That is a coaching problem before it is a talent problem, and Ohio State has the talent.
For Texas, the win is the kind that reshapes a season on both ends, the players' belief and everyone else's read on them. Comebacks that size are rare enough that they become the organizing story of a year whether the team wants them to or not.
What I'm watching next is simple: the FPI top 25 a month from now, once the preseason weight has burned off and the results are doing the talking. If Ohio State is still at or near the top then, this poll was early and right. If the Buckeyes slide, the poll was early and early. Either way, the shouting today is aimed at the wrong target.
