Every FBS team. Same field, same pace.

The All-Takes
Power Rating

Like other college football ratings, the All-Takes power ratings answer a simple question: if two teams met on a neutral field tomorrow, how much better would one be? What makes All-Takes different is the transparency of the answer.

138FBS teams rated
815Weeks 1–4 games tested
74.0%winners picked
+2.9 ptsversus preseason FEI

How to read it

Points, not mystery units.

Ohio State at +29.4 and an average team at 0 means roughly a 29-point neutral-field gap under normal game conditions. To compare any two teams, subtract the lower rating from the higher one.

What goes into a pre-season rating? Each rating starts with how well the team played, given the quality of its opponents and location of each game. It then updates for the 2026 roster and staff changes—specifically quarterback, offensive line, returning production, recruiting, transfers, NFL departures and coaching—and for how the team finished 2025. Lastly, it uses early-season betting lines selectively as a modest adjustment.

These ratings compare teams on a neutral field under normal game conditions. A prediction for an actual game separately adds home field and how the teams match up. It also estimates the likely number of competitive possessions from each team's tempo, drive length, efficiency, explosiveness, turnovers and play style. That estimate makes a small Week 1 adjustment in projected non-blowouts, but it cannot choose a side on its own. One-game effects do not become permanent team strength.

Late-2025 Form gives a team some credit back for games its starting quarterback missed. It barely changes the Week 1 ratings, but helps keep a short-term injury from looking like a lasting decline in team quality.

Positive factors help a rating. Negative factors pull it down. Every factor shown below adds exactly to the published number.

The complete board

Every FBS team, ranked and explained.

Select a team to see how four pieces add up: its starting point, 2026 roster and staff, Late-2025 Form and a limited check against the Week 1 betting line.

138 shown
Rank / team / other modelsLargest 2026 movesMedian winsRating

Win totals cover the currently scheduled regular season only. Each matchup is played 1,000 times, with game uncertainty resampled from 815 frozen 2022–25 preseason predictions (13.02-point average absolute error). “At” is the share landing exactly on the displayed integer median; below, at and above always sum to 100%.

What changed—and who disagrees

Three arguments hiding inside the board.

Rank gaps are the starting point, not the conclusion. These are the five largest moves in each comparison, followed by the clearest football explanation the underlying inputs suggest.

01 · Year over year

The five biggest moves since the end of 2025

Absolute rank change among returning 2026 FBS teams, comparing the final 2025 time-weighted All-Takes state with the 2026 board.

  1. North Texas#39 → #85 · ↓ 46

    The Week 1 market check removes 4.23 points, and a 31% retained-roster share adds another major drag. Quarterback help is not enough to offset the roster, transfer and coaching deductions.

  2. Toledo#41 → #80 · ↓ 39

    A 4.18-point Week 1 market correction drives most of the fall. Lower talent continuity and a coaching-transition penalty deepen it.

  3. James Madison#27 → #62 · ↓ 35

    The 2025 starting point remains strong, but only 42% of the roster is retained. Roster loss, portal/NFL departures, a coaching change and the Week 1 line all point down.

  4. Rutgers#66 → #38 · ↑ 28

    The context-filtered Week 1 market signal adds 3.77 points. A positive quarterback projection and net portal/draft contribution help preserve the move despite ordinary roster continuity.

  5. Old Dominion#53 → #79 · ↓ 26

    The largest deduction is 2.53 points for returning-roster quality, with smaller talent and portal/NFL losses. There is no lined FBS Week 1 game to provide a market counterweight.

02 · ESPN SP+

Where All-Takes and SP+ disagree most

Absolute difference between the final 2026 preseason ranks.

  1. Oklahoma StateAT #93 · SP+ #39
    All-Takes is 54 spots lower

    SP+ is much more willing to buy the program-level rebound. All-Takes starts from a deeply negative 2025 baseline, sees only 22% of the roster retained and gets another negative signal from the Week 1 line.

  2. TulaneAT #48 · SP+ #87
    All-Takes is 39 spots higher

    All-Takes carries forward a strong 2025 level and late-season form, then gets a 1.65-point lift from the Week 1 market. SP+ appears more sensitive to the Green Wave's roster and coaching turnover.

  3. North CarolinaAT #89 · SP+ #56
    All-Takes is 33 spots lower

    All-Takes sees a weak 2025 starting point and a 3.08-point negative market check. SP+ lands closer to the middle of FBS, likely retaining more faith in the Tar Heels' underlying talent and program baseline.

  4. Sacramento StateAT #100 · SP+ #131
    All-Takes is 31 spots higher

    Both systems are cautious with a new FBS member, but All-Takes begins from an FCS performance seed instead of an almost-bottom-of-FBS prior. A negative Week 1 market signal keeps this from being an even larger gap.

  5. Iowa StateAT #39 · SP+ #69
    All-Takes is 30 spots higher

    All-Takes gives Iowa State substantial credit for its 2025 level, late form and quarterback projection. SP+ appears to punish the Cyclones' 35.5% retained-roster share and broader talent losses more aggressively.

Final 2026 SP+ source ↗
03 · BCF Toys FEI

Where All-Takes and FEI disagree most

Absolute difference between the 2026 preseason ranks.

  1. North Dakota StateAT #53 · FEI #131
    All-Takes is 78 spots higher

    FEI gives new FBS teams a bottom-tier -1.00 preseason prior because they lack five years of FBS possession history. All-Takes uses North Dakota State's FCS performance seed and a strong Week 1 market confirmation.

  2. Sacramento StateAT #100 · FEI #132
    All-Takes is 32 spots higher

    FEI applies the same history-limited -1.00 prior to Sacramento State. All-Takes also penalizes the roster and Week 1 line, but its FCS seed still leaves the Hornets well above FEI's placement.

  3. RutgersAT #38 · FEI #67
    All-Takes is 29 spots higher

    FEI's five-year possession history remains skeptical. All-Takes reacts more strongly to Rutgers' current quarterback/portal picture and, above all, a 3.77-point context-filtered Week 1 market signal.

  4. James MadisonAT #62 · FEI #38
    All-Takes is 24 spots lower

    FEI continues to reward James Madison's multi-year possession efficiency. All-Takes responds more aggressively to the current 42% retained-roster share, coaching change, departures and negative Week 1 line.

  5. North TexasAT #85 · FEI #62
    All-Takes is 23 spots lower

    FEI's longer history changes slowly. All-Takes sees immediate 2026 risk: a 31% retained-roster share, coaching and transfer deductions, plus a 4.23-point negative Week 1 market correction.

2026 preseason FEI source ↗

Past results

Does it actually work?

To keep the winner test honest, each All-Takes rating was rebuilt using only seasons that had already been played, locked before kickoff, and scored over Weeks 1–4. Preseason FEI got the same treatment on the same 815 FBS games.

All-Takes74.0%

603 of 815 winners picked

Published preseason FEI71.0%

579 of 815 winners picked

Net result+24

more correct picks for All-Takes on the common sample

SeasonGamesAll-TakesPreseason FEI
202221073.8% (155)70.5% (148)
202321677.3% (167)73.1% (158)
202419473.2% (142)69.1% (134)
202519571.3% (139)71.3% (139)

Against the spread

Has it beaten the spread? Historically, yes.

The policy does not bet every game. From 2017–2025, it selected 588 sides and went 337–251. It finished above the standard −110 break-even rate in 8 of 9 seasons.

Selected-bet record57.3%

337 wins · 251 losses

−110 break-even52.4%

the percentage a bettor must beat

Flat-stake return+9.4%

historical return per unit risked at −110

Winning seasons8 of 9

seasons above the break-even rate

A stricter check—where every season was judged using rules learned only from earlier seasons—went 269–203 (57.0%) across 472 selections. These are historical research results against a multi-provider line proxy, not a live sportsbook record or a guarantee.

See the current ATS card and every reason

Familiar national models

Against FPI and SP+, it is essentially a tie.

Over 2024–2025, All-Takes and SP+ each picked 281 of 389 winners; FPI picked 280 and preseason FEI picked 273. In 2025 All-Takes trailed FPI by one winner and SP+ by four; its stronger 2024 result erased that gap. Two seasons are informative, not final.

All-Takes72.2%

281 of 389 winners

ESPN preseason FPI72.0%

280 of 389 winners

ESPN preseason SP+72.2%

281 of 389 winners

Published preseason FEI70.2%

273 of 389 winners

The changing pregame line did best at picking outright winners: 290 of 389 (74.6%). That is a different question from ATS. Winner accuracy asks who won; ATS asks whether the selected team covered the spread.

Source snapshots: ESPN FPI, 2024 · ESPN SP+, August 13, 2024 · ESPN FPI, August 15, 2025 · ESPN SP+, August 13, 2025

The main ingredients

What the model considers.

Hundreds of details feed the model, but they answer four familiar football questions.

01

How good have they been?

Past results are adjusted for the strength of the opposition. Offense, defense and special teams all matter, while one great or terrible game cannot erase the larger record.

02

Who is on the team now?

The model checks who returned, arrived and left, with special attention to quarterback, offensive line, top-end talent, transfers, NFL departures and coaching changes.

03

What was real—and what was noise?

Recent form matters, but weather, turnovers and late scores in decided games receive less weight. Strong play against strong opponents receives more.

04

What changes in this matchup?

For a game prediction, the model considers location, how the teams match up and more of the latest betting line. It estimates expected possessions and makes a small pace adjustment in projected non-blowouts. The rating keeps only the limited market check described above; game-only adjustments do not permanently redefine either team.

What can still change

Updated as rosters, depth charts and lines change.

These are preseason ratings. Injuries, quarterback decisions, player roles, late transfers and moving betting lines can change the picture. Historical quarterback absences are lightly corrected in the carried rating; current-game injury adjustments wait for verified availability. When reliable information is missing, the model leans more on the team's track record instead of guessing. Every input to a prediction must have been available before kickoff.

See Week 1 ATS recommendations, decomposed by reason Ratings snapshot: August 9, 2026.