Who Holds the Top 25? Six Weeks of Ranking Capital
Ranking Capital through Week 7 Polls
Six weeks of polls, one method for every conference, and what a loss, a ranked win and a record really do to a team's place in the poll.
· Launch article · Ranking Capital through Week 7 Polls · Polls: AP and Coaches, through week 7 · Data: game results through week 6
Alabama beat Mississippi State by 33 points, and Mississippi State dropped one spot in the AP poll. Ohio State beat Iowa by 17, and Iowa dropped six. The poll didn't change its rules between those two games, so we went looking for what it actually does with a loss. First, a way to see how much of the poll each conference holds.
What we measured
Ranking Capital (RC) gives a ranked team points for its place in a poll: No. 1 earns 25, No. 25 earns 1, and an unranked team earns 0. We add those up by conference, so a conference's total shows how much of the Top 25 it holds. Official RC averages the AP and Coaches polls; the CFP rankings join once they are published.
RC is a snapshot of one week's polls, and it does not decide anything. It can't prove conference quality, media bias, voter intent, or that a conference deserves more ranked teams. It shows where the poll's weight sits, week by week, so we can ask better questions about it. Polls run through week 7, which is published before week 7's games, so game results here cover weeks 1 to 6.
The season so far, week by week
Official RC by conference, from the preseason poll (week 1) through week 7.
| Conference | Wk 1 | Wk 2 | Wk 3 | Wk 4 | Wk 5 | Wk 6 | Wk 7 |
|---|---|---|---|---|---|---|---|
| SEC | 131.5 | 138 | 139 | 133 | 137.5 | 141.5 | 128.5 |
| Big Ten | 111.5 | 101 | 89.5 | 93.5 | 78 | 73 | 76.5 |
| Big 12 | 33 | 34 | 40 | 44 | 54 | 57.5 | 63.5 |
| ACC | 28 | 30 | 33.5 | 31.5 | 30.5 | 28.5 | 30 |
| Independent | 21.5 | 22.5 | 23 | 23 | 23 | 23 | 24 |
| Pac-12 | 0 | 0 | 0 | 0 | 2 | 1.5 | 2.5 |
Listed by week 1 RC, highest first.
The American, Conference USA, Mid-American, Mountain West and Sun Belt hold 0 in every week of both polls.
Two things worth a look:
- The Big Ten slid, the Big 12 climbed. The Big Ten went from 111.5 to 76.5 across the season, while the Big 12 nearly doubled, from 33 to 63.5.
- The SEC held steady near the top. It stayed between about 128 and 142, and its week 7 figure is its lowest of the season.
Lose by 33, drop one. Lose by 17, drop six.
Back to the question that opened this piece: what does a loss cost a ranked team? Alabama (#7) beat Mississippi State (#16) 56-23. Mississippi State went to #17 in the next AP poll. Ohio State (#5) beat Iowa (#14) 31-14. Iowa went to #20. The smaller margin, against the higher-ranked winner, cost six spots.
The explanation is in the traffic around them. Six teams passed Iowa, and two of those, Oregon (15 to 13) and Oklahoma State (19 to 18), didn't even play. Missouri jumped from #25 to #14 by beating #8 Florida 45-17. Mississippi State was passed by only two teams, while Iowa and Kentucky fell below it. Kentucky fell out of the poll entirely after a 35-34 win over an unranked team.
That is not a one-off. Among this season's AP losses by ranked teams to top-10 opponents, drops ran from 0 to 6 spots, and one team fell out. Across all 26 AP losses by ranked teams this season, the size of the margin doesn't line up with the size of the fall: teams that stayed ranked dropped an average of 5.3 spots after losing by 1 to 7 points, 8.8 after losing by 8 to 14, and 5.0 after losing by 15 or more. A loss moves a team less than who else lost, who had a bye, and who was waiting in line.
Two ranked teams didn't drop at all after losing: Louisville (ACC) stayed #24 after a 41-38 loss to #9 Ole Miss, and Alabama (SEC) stayed #6 after a 20-17 loss to #2 Georgia. Here is every AP loss by a ranked team through week 6, by conference:
| Conference | Losses | Fell out | Avg drop, teams still ranked | Largest drop |
|---|---|---|---|---|
| SEC | 13 | 2 | 6.45 | 14 |
| Big Ten | 7 | 3 | 8.0 | 15 |
| ACC | 5 | 3 | 3.0 | 6 |
| Big 12 | 1 | 0 | 3.0 | 3 |
Listed by number of losses, most first.
The Big Ten's ranked losers fell the furthest, and 3 of its 7 fell out. The ACC's still-ranked teams moved the least, but 3 of its 5 fell out. Those averages leave out the teams that fell out, which are the biggest penalties. The samples are small (the Big 12 has one loss), so this describes the first six weeks and doesn't show a pattern.
Two losses by a No. 1 team show how much the surrounding teams matter. In week 2, #1 Ohio State lost 24-23 at Texas, then ranked #4 in the AP poll, and fell to #6. In week 6, #1 Texas lost 27-24 to an unranked Oklahoma and fell to #7 in the AP poll (#6 in the Coaches Poll). One loss was to a top-5 team and one to an unranked team, yet both ended up in nearly the same place, because a #1 team that loses drops below the teams that were next in line.
Five teams passed Ohio State. Six ended up ahead of Texas: Georgia, Notre Dame, Miami, Ohio State and Indiana moved up, and Alabama stayed at #6 after losing to #2 Georgia. Most of the teams that moved up past either team won big against unranked opponents. Whether either drop was deserved is a separate question. The data shows where each team landed and who stood around it, not how voters weighed the two games.
Does a ranked win hold its value?
Beating a ranked team feels like a résumé line. Whether it stays one depends on what that team does next.
Most ranked teams do hold on: 92% of AP-ranked teams are still ranked in the next poll, and 80% three polls later. But teams that just lost while ranked are a different group. Of the 20 AP-ranked teams that lost and have a poll two weeks later, 8 were out of the poll by then, and 7 of the 17 with a poll three weeks later. That is the pool a win over a "ranked team" is drawn from.
Florida is the season's best example. It beat #4 Ole Miss and jumped from #21 to #8. It then lost to Missouri (#25) and to South Carolina (unranked), fell to #16, and dropped out. The teams around it did not follow one pattern: Ole Miss slid from #4 to #9, Missouri rose from #25 to #14 and was still #13 two polls later, South Carolina stayed unranked at 3-3, and Auburn, which lost to Florida earlier, stayed unranked.
Conferences differ a little here. In the AP, 52 of 56 ranked SEC teams (93%) were still ranked in the next poll, against 36 of 41 for the Big Ten (88%) and 16 of 19 for the ACC (84%). The ACC's rate is lower, but with 19 team-weeks behind it, a gap of this size is within what a few results can produce, so we don't read it as a pattern.
Last season as a mirror
To compare equal stretches, look at the AP poll from the preseason poll through week 7. In 2025, 14 teams entered the poll and 14 left it over those six changes, and 28 moves were 5 spots or more. In 2026 it is 12 in, 12 out and 19 big moves. The total rank change was 380 in 2025 and 306 in 2026. The Coaches Poll shows the same: 15 in and out in 2025, 13 in 2026. By this measure the churn this year is a little lower than last year, not higher.
Is 2026 following the same path? Partly:
| Measure | 2025 | 2026 |
|---|---|---|
| AP-ranked teams still ranked at +1, +2, +3 polls | 89%, 82%, 78% | 92%, 87%, 80% |
| AP-ranked teams that lost and were out of the next poll | 14 of 29 | 8 of 26 |
| SEC RC, week 1 to week 7 | 127.5 to 148 (+16.1%) | 131.5 to 128.5 (−2.3%) |
| Big Ten RC, week 1 to week 7 | 98.5 to 89.5 (−9.1%) | 111.5 to 76.5 (−31.4%) |
| Big 12 RC, week 1 to week 7 | 31 to 33 (+6.5%) | 33 to 63.5 (+92.4%) |
| ACC RC, week 1 to week 7 | 47 to 41.5 (−11.7%) | 28 to 30 (+7.1%) |
| Independent RC, week 1 to week 7 | 20.5 to 10 (−51.2%) | 21.5 to 24 (+11.6%) |
| Pac-12 RC, week 1 to week 7 | 0 to 0 | 0 to 2.5 (n/a) |
| American RC, week 1 to week 7 | 0 to 3 (n/a) | 0 to 0 |
| Conference USA RC, week 1 to week 7 | 0 to 0 | 0 to 0 |
| Mid-American RC, week 1 to week 7 | 0 to 0 | 0 to 0 |
| Mountain West RC, week 1 to week 7 | 1 to 0 (−100%) | 0 to 0 |
| Sun Belt RC, week 1 to week 7 | 0 to 0 | 0 to 0 |
RC rows are listed by 2026 week 7 RC, highest first; conferences at 0 are alphabetical.
Survival rates look alike. The conference picture doesn't: last year the SEC's weight grew through week 7, and this year it hasn't.
Does starting with more ranked teams change the picture?
A conference that starts with ten ranked teams has more chances to rack up ranked wins and ranked losses than one that starts with four. That is a fair objection to any ranked-win count, so we measured the chances themselves.
| Conference | Ranked in preseason AP, 2025 | Ranked in preseason AP, 2026 | Games vs a ranked opponent, 2025 | Games vs a ranked opponent, 2026 | Record in those games (win %), 2025 | Record in those games (win %), 2026 |
|---|---|---|---|---|---|---|
| SEC | 10 | 9 | 27 | 33 | 11-16 (.407) | 14-19 (.424) |
| Big Ten | 6 | 8 | 20 | 23 | 5-15 (.250) | 6-17 (.261) |
| Big 12 | 4 | 4 | 12 | 16 | 5-7 (.417) | 3-13 (.188) |
| ACC | 3 | 3 | 13 | 15 | 6-7 (.462) | 3-12 (.200) |
| American | 0 | 0 | 8 | 6 | 2-6 (.250) | 0-6 (.000) |
| Conference USA | 0 | 0 | 5 | 6 | 0-5 (.000) | 0-6 (.000) |
| Mid-American | 0 | 0 | 7 | 5 | 0-7 (.000) | 0-5 (.000) |
| Mountain West | 1 | 0 | 7 | 5 | 0-7 (.000) | 0-5 (.000) |
| Pac-12 | 0 | 0 | 2 | 11 | 0-2 (.000) | 0-11 (.000) |
| Sun Belt | 0 | 0 | 9 | 5 | 0-9 (.000) | 0-5 (.000) |
| Independent | 1 | 1 | 2 | 0 | 0-2 (.000) | 0-0 |
The four conferences with the most 2026 preseason ranked teams come first, in that order; the rest are alphabetical, with the Independents last.
Both seasons cover games through week 6. The Pac-12 had 2 teams in 2025 and 8 in 2026, so its row isn't like for like. The SEC had more ranked-opponent games than any other conference in both seasons (27 in 2025, 33 in 2026), and the Big Ten, Big 12 and ACC had fewer. When two ranked teams from the same conference play, one of them takes a ranked win and the other a ranked loss, so a conference with many ranked teams sees both pile up.
Now the promotion question: if a conference had started with more ranked teams, would its winners have been promoted as much? We can't re-run the season, so we compared what we can. The chart below shows how far a winner moved up in the next AP poll, by the rank of the team it beat.
Per win, the rank of the team you beat matters most: the gain shrinks as the opponent's rank drops, in both seasons. Per season, a conference with more ranked teams has more of those valuable wins available, and more losses to trade back. What the data can't do is test whether another conference would be promoted as much if it started with more ranked teams. That takes more than two seasons, and most conference-by-opponent-rank groups have fewer than five wins.
What if only the final poll counted?
Here is a stricter test. Take the 25 teams in the final 2025 AP poll (a full-season test, since 2026 has no final poll yet) and count only those as ranked teams for the whole season. A win over any team ranked at kickoff that did not finish in the top 25 is vacated: LSU, which was ranked early and finished 7-6, no longer counts for Ole Miss, Vanderbilt or Texas A&M.
Of the 85 wins over teams ranked at kickoff, 49 are vacated. The SEC's 29 become 13, the Big Ten's 14 become 7, and 15 wins over eventual final-poll teams that were unranked at the time are added.
| Conference | Final-poll teams | Games vs final-poll teams (regular season) | Record | Record vs other conferences |
|---|---|---|---|---|
| SEC | 7 | 64 | 14-50 (.219) | 3-3 (.500) |
| Big Ten | 6 | 63 | 10-53 (.159) | 1-6 (.143) |
| Big 12 | 5 | 49 | 13-36 (.265) | 1-1 (.500) |
| American | 3 | 35 | 4-31 (.114) | 0-8 (.000) |
| ACC | 2 | 32 | 8-24 (.250) | 3-11 (.214) |
| Sun Belt | 1 | 18 | 0-18 (.000) | 0-9 (.000) |
Listed by number of final-poll teams, most first. The Independents' one final-poll team, Notre Dame, played 4 games against final-poll teams and went 2-2; it is not shown.
Counting each conference's games against final-poll teams per member team, the SEC faced the most: 4.0 games per team, against 3.5 for the Big Ten, 3.1 for the Big 12 and 1.9 for the ACC. In the postseason against final-poll teams, the ACC went 3-2 (.600), the Big Ten 6-5 (.545), and the SEC 4-8 and the Big 12 1-2 each won a third of their games (.333).
Two cautions. First, a final-poll list is built from results, so teams that win a lot are more likely to be on it, and games against them are mostly losses for everyone. Second, the cross-conference records here are 2 to 14 games each. The same test can't be run for 2026 until its final poll exists.
The conference ledger
Every conference, shown the same way: its record against FBS teams from other conferences this season (games through week 6), the average margin, and its share of the poll in week 7.
| Conference | Record vs other FBS conferences (win %) | Avg margin | Week 7 RC |
|---|---|---|---|
| SEC | 28-2 (.933) | +24.70 | 128.5 |
| Big Ten | 30-11 (.732) | +15.98 | 76.5 |
| Big 12 | 23-8 (.742) | +11.26 | 63.5 |
| ACC | 18-12 (.600) | +2.73 | 30 |
| Pac-12 | 9-16 (.360) | -4.36 | 2.5 |
| Mountain West | 10-16 (.385) | -6.12 | 0 |
| American | 15-23 (.395) | -6.97 | 0 |
| Sun Belt | 14-26 (.350) | -8.72 | 0 |
| Conference USA | 6-24 (.200) | -13.50 | 0 |
| Mid-American | 7-27 (.206) | -22.12 | 0 |
| FBS Independents | 8-3 (.727) | +19.27 | 24 |
Listed by average margin, highest first, except the FBS Independents, who are shown last.
Conferences with big margins also hold the most poll weight this season, and the conferences with negative margins hold almost none. The Independents' 8-3 comes from 11 games, and the Pac-12 holds 2.5 despite a 9-16 record, so small samples are part of this picture.
Same weeks last year (2025, games through week 6): the SEC went 34-6 (.850) against other FBS conferences (+21.0 per game) and is 28-2 (.933, +24.7) this year; the Big 12 went 22-9 (.710, +10.3) and is 23-8 (.742, +11.3); the Big Ten went 30-9 (.769, +18.6) and is 30-11 (.732, +16.0); the ACC went 20-16 (.556, +2.8) and is 18-12 (.600, +2.7). They are listed by this year's win percentage. The Pac-12 and Conference USA have different team counts in the two seasons, so their rows aren't a like-for-like comparison.
We aren't reading a verdict into those numbers. A conference can hold a large share of the poll and have the results to match, and samples of about 10 to 41 games can swing either direction. If you want to argue it, the ledger is the place to start.
What we can't say
- Small samples. Six weeks of games is about 10 to 41 cross-conference games per conference. One upset moves an average margin a lot.
- Only games in our database. We have the 2025 and 2026 seasons, nothing earlier, so we make no claims about longer trends.
- No schedule-strength adjustment. A 28-2 record doesn't say who the opponents were.
- No claim about motive. We can show how voters' lists changed. We can't show why.
- 2025 conference labels come from the data. Only the 2026 season has an approved conference membership list, so 2025 groupings use the conference listed in the source data for each game.
- Games run through week 6; the polls run through week 7.
Use it yourself
Every number here comes from pages you can open. The Rankings page carries the weekly RC panel, the RC page has the conference table and the season-at-a-glance chart, and each team page shows its poll history. Two more studies are coming next: PIP and PvE.