Polls, Committees, and Chaos: How College Football Rankings Work

Every August, college football produces two rankings, one committee, and roughly ten million arguments. The 2026 preseason AP Top 25 just dropped — Ohio State at 1, defending champion Indiana all the way down at 6 — and if you have ever wondered who actually decides these things, what the numbers are for, and why a team can win the national title in January and start the next season behind five teams that did not, this is your field guide. The system is genuinely strange. It is also, once you see the machinery, one of the sport's best inventions.

Start with the oldest instrument in the toolbox. The Associated Press poll has been ranking college football teams since 1936, which means it predates the NFL's television era, the facemask, and most of the stadiums it covers. The mechanics are simple: 65 voters — sportswriters and broadcasters who cover the sport across the country — each submit a ballot ranking their top 25 teams. A first-place vote is worth 25 points, a second-place vote 24, and so on down to a single point for 25th. Add up the points, sort the totals, and you have the poll.

When you hear that a team is "No. 1 in the AP poll," you are hearing the aggregate opinion of 65 professional watchers, updated weekly, every one of them working from the same incomplete evidence as the rest of us.

The Coaches poll runs on the same arithmetic with a different electorate: a panel of active FBS head coaches submits ballots, and the totals stack the same way. In practice the two polls track each other closely — this year's preseason editions are nearly identical — but the differences are instructive. Writers watch everything; coaches watch their own film and trust their networks. When the polls disagree about a team, it usually means the eye test and the coaching grapevine are seeing different things, and that gap is often where the season's surprises hide.

Here is the part that trips up new fans: for most of college football's history, these polls did not just measure the championship race. They were the championship. There was no tournament. The final AP ballot, taken after the bowls, crowned the national champion, and the Coaches poll crowned its own — and they did not always agree. As recently as 1997, the sport split its title: Michigan finished No. 1 with the writers while Nebraska finished No. 1 with the coaches, and both programs claim that championship to this day. Imagine the NBA ending its season with two trophies awarded by two separate press conferences. That was college football, on purpose, for six decades.

The sport spent a generation trying to engineer the argument out of existence. The BCS arrived in 1998, blending polls with computer ratings to force the top two teams into a single title game — progress, but the formula itself became the controversy, because a machine ranking teams invites everyone to audit the machine. So in 2014 the sport tried something more human: the College Football Playoff, with a 13-member selection committee — athletic directors, former coaches, former players — charged with picking and seeding a four-team bracket. The polls were formally retired from the deciding business. They became what they are now: the sport's running conversation, influential but not binding.

Then the bracket grew. Beginning with the 2024 season, the playoff expanded to 12 teams, and the whole sport reorganized around the new math. Under the current format, the five highest-ranked conference champions earn automatic bids, the remaining spots go to the committee's highest-ranked at-large teams, the top seeds earn first-round byes, and the opening round is played on campus — home stadiums, December weather, tickets in the hands of actual students.

Ohio State won the first edition of the expanded bracket in January 2025; Indiana won the second this past January. Two years, two champions, and a December tradition — playoff football in home stadiums — that already feels permanent. The committee meets late in the season, releases its own rankings weekly, and those rankings, not the AP's, decide who gets in and where they are seeded.

Which raises the reasonable question: if a committee controls the bracket, why does anyone care where the AP puts teams in August? The honest answer is path dependence. Polls move incrementally. Voters do not rebuild their ballots from scratch each week; they adjust — winners drift up, losers drift down, and teams mostly climb by beating somebody ranked above them. That means your starting position shapes your whole route. A team that opens at No. 6 and wins every week will be in the playoff conversation by mid-October. A team that opens unranked and wins the exact same games has to spend a month proving it belongs on the ballot at all before it can start climbing. August is where the sport sets everyone's base camp.

The preseason poll also quietly writes the season's script. Rankings determine which September games get flagged as showdowns, which results count as "ranked wins" in every argument that follows, and which upsets register as earthquakes. When No. 24 Louisville visits No. 9 Ole Miss on the season's first weekend, that game carries weight precisely because two poll numbers are attached to it. Strip the numbers away and it is a good game. Attach them and it is evidence — the kind the committee will hear about in November when resumes get compared.

The 2026 poll is a perfect worked example of all this machinery humming at once. Ohio State opens at No. 1 because the voters treat recent championship infrastructure as the safest August bet. Indiana opens at No. 6 despite being the defending champ because preseason ballots price in roster turnover and the sheer depth of the contender class — five teams the voters simply believe in enormously. Nobody thinks the Hoosiers got worse at winning in January. The voters think the mountain got taller. If Indiana is as good as its banner, the poll will correct itself one Saturday at a time, which is exactly how the instrument is designed to work.

One more thing the rankings do, and it might be the most important: they metabolize hope. Sixty-five ballots in August are 65 educated guesses, and everyone involved knows a chunk of them will look silly by Halloween. That is not a bug in the system. The preseason poll exists to be argued with, corrected, and eventually overwritten by actual football — a hypothesis the season spends fourteen weeks testing. The committee will render the verdicts that matter in December. Until then, the polls give every fan base a number to defend or a slight to avenge, which is the cheapest, most renewable fuel in sports.

So read the August rankings the way a coach reads a depth chart in spring: a snapshot, not a destiny. The 2026 edition says Ohio State is the standard, the champion is sixth, and 23 other teams are officially in the argument. The season starts settling it August 29. The final poll — the one that gets remembered — does not exist yet, and every team in the country still holds a pen.