Michigan Is The Only Team That Left The AP Top 25, And It Won

There were 25 teams in the Associated Press preseason poll and there are 25 teams in the one that came out on Tuesday afternoon. Exactly one name left the list.

It is the name of a team that won its game.

Michigan opened the season 16th in the AP preseason ballot with 718 points, beat Western Michigan 13-12 at Michigan Stadium, and is not in this week's top 25. Virginia moved in at 25th and took the vacated place. Every other team from the preseason list is still on it.

That includes Louisville, which is the only preseason-ranked team that lost.

What the ballot actually did#

Louisville went to Oxford and lost 41-38 to Ole Miss, ranked ninth, on a walk-off field goal. It stayed in the poll, at 24th, one spot below where it started. Michigan won and left.

The Associated Press flagged both halves of that, and both are records in its own files. Of Michigan it wrote that this is "the first time since at least 2008 that a team ranked so high was bounced after a win," the nearest thing to a precedent being Kansas State, 17th to unranked last season after surviving a Football Championship Subdivision opponent. Of Louisville it wrote that the Cardinals "became the lowest-ranked team since at least 2008 to stay in the Top 25 after a loss." Two eighteen-year firsts, pointing in opposite directions, in one ballot, and the two teams that produced them are the two in this paragraph.

The headline AP put on the poll names the thing this beat has to name with it: Michigan fell out "after second-chance Hail Mary." The win was 13-12 and it required a replay official putting one second back on the clock after a first Hail Mary fell incomplete, and by the time voters filed on Tuesday the Mid-American Conference had formally asked the NCAA to award the game to Western Michigan instead. The NCAA said the result stands. Whatever a voter thought of that sequence, no voter ranking Michigan on Tuesday was ranking a clean one-point win, and pretending the ballot turned only on the margin would be reading the poll with one eye shut.

The gap between those two outcomes is the whole finding, and it is not a complaint. Voters were not asked to rank results. They were asked to rank teams, and 69 first-place votes went out this week, one to a ballot, which is the size of the electorate doing the asking. What that electorate saw was a three-point road loss at a top-ten team and a one-point home win over a Mid-American Conference team, and it decided the first was better evidence than the second.

The margin is the part that carries. Twenty-five preseason-ranked teams played last week. Michigan's one-point win is the narrowest result any of them produced, in either direction. Ole Miss and SMU each won by three, Oregon by seven, Houston by 13. Then the list jumps: Notre Dame by 28, Iowa by 40, Penn State by 45, Oklahoma by 51, Ohio State by 53, Georgia by 60.

The coaches' poll reached a softer version of the same verdict. Michigan started 16th there with 719 points and is 24th now with 164, which is 555 points of support gone in one week without losing. Clemson, which lost 51-10 at LSU, fell out of the coaches' poll entirely. It was never in the AP's 25 to fall out of.

What a week-two poll is actually worth#

Very little, and saying so is not a defence of Michigan.

A poll released after one week is a preseason guess with a single result attached to it. The preseason ballot was written in July by people who had watched no football, and what any voter had by Tuesday afternoon was one Saturday and three days of argument about it. Michigan's drop is not a measurement of how good this team is. It is a group of voters revising a July estimate downward after the first data point disagreed with it, and after that data point spent three days being disputed in public.

The same caution runs the other way. Georgia beating Tennessee State by 60 established nothing either. Six of the top 10 played a team from outside the top division or the bottom of the Group of Five. Almost nobody learned anything about themselves last weekend, and the poll pretends otherwise because the poll comes out weekly whether or not there is anything to put in it.

What is real is the schedule.

The schedule is the argument#

Michigan has 11 games left. Six of them are against teams in the poll it just dropped out of: Oklahoma 11th, Penn State 16th, Iowa 21st, Indiana fifth, Oregon sixth, Ohio State first.

Three of those six are in the current AP top six. Only four other teams in the country play three of the top six this season, and three of them are in the same conference: Nebraska, Northwestern, USC, and Ohio State itself.

The other five games, against UTEP, Minnesota, Rutgers, Michigan State and UCLA, are against unranked teams and offer a Michigan resume nothing at all. That is the structural fact of this season. There are six games that can build a case and five that can only damage one.

Michigan gets the first of the six on Saturday. Oklahoma is at Michigan Stadium at noon Eastern on Fox, a sold-out Maize Out, per Michigan's own schedule page opened this morning. Oklahoma opened its season by beating UTEP 51-0.

Michigan plays UTEP on September 19. That is a common opponent arriving in 10 days, and it is the cleanest early measuring stick this schedule offers, whatever happens on Saturday.

The playoff math, stated honestly#

There is no College Football Playoff ranking right now. The selection committee does not publish until November, and any number attached to Michigan before then is somebody's projection rather than a ranking. Michigan's actual national standing today is one line long: unranked in the AP poll, 24th in the coaches' poll, 1-0.

What the schedule says about the path is arithmetic rather than opinion. A team whose only ranked opponents are its last six chances has to beat ranked teams to have a case, because it cannot build one out of UTEP and Rutgers. Michigan does not need to be judged well. It needs the six results, or enough of them, and the first one is in four days.

The counter-argument to all of the above is that a team can be badly underrated in September and correct the record itself. That is true and it happens most years. It also requires beating somebody, which Michigan has not yet done.

What carries forward#

Andrew Babalola, the sixth offensive lineman in the package Jason Beck brought from Utah, did not play in the opener. Kyle Whittingham ruled him out on the Monday before the game and said afterwards that he remains unavailable. Michigan has published no injury report and is not required to, and the participation report for each game remains the only document that says who was on the field.

That is the thread the weekly game coverage cannot hold, because each Saturday only sees itself. Six ranked opponents is a demand on a roster's depth, not just on its best 11 players, and Michigan is starting that stretch with its own coach describing a lineman as still on the shelf.

The ballot will keep moving weekly. The schedule will not.

Sources#

• Eric Olson, the Associated Press, "LSU cracks top 10, Oregon slips and Wolverines fall out of AP Top 25 after second-chance Hail Mary", published 2:00 p.m. Eastern on September 8, opened September 9, for the Week 1 poll, both of the since-2008 records and the Louisville result.
• Associated Press Top 25 and USA Today Coaches Poll, preseason and Week 1 releases, and the 2026 Michigan and full FBS schedules and results, via the CollegeFootballData API, read 5:26 a.m. Pacific on September 9. The poll's release day is taken from the AP article above rather than from the feed's week label.
• Michigan's 2026 football schedule, mgoblue.com, opened September 9, for Saturday's kickoff time, network and Maize Out designation.
• Michigan Athletics postgame notes and Michigan Monday release, September 5 and September 7, for the final score and Oklahoma's ranking as Michigan's own site carried it.