The Poll.ai

Connectivity report · 2025 week 4

Nobody knows anything yet. Here is exactly how much.

The poll opens in Week 5. That date was published before the season and does not move. This is week 4, and what follows is a description of what is not yet knowable.

Every other poll is publishing a top 25 this week. It is built from reputation, because there is nothing else to build it from yet. This page is the alternative: a measurement of how little the schedule has told the model, published in the open, with the date the real poll opens fixed in advance.

The schedule graph

Abilene ChristianAdrianAir ForceAkronAlabamaAlabama A&MAlabama StateAlbany State GAAlcorn StateAllenAmerican InternationalApp StateArizonaArizona StateArkansasArkansas StateArkansas-Pine BluffArmyAuburnAustin PeayBLUEFIELDBYUBall StateBaylorBentleyBethune-CookmanBoise StateBoston CollegeBowie StateBowling GreenBrownBryantBucknellBuffaloButlerCal PolyCaliforniaCampbellCentral ArkansasCentral ConnecticutCentral MichiganCentral State (OH)Central WashingtonChadron StCharleston SouthernCharlotteChattanoogaCincinnatiClemsonCoastal CarolinaColgateColoradoColorado StateColumbiaCornellCumberland (TN)DartmouthDavidsonDaytonDelawareDelaware StateDrakeDukeDuquesneERSKINEEast CarolinaEast Tennessee StateEast Texas A&MEastern IllinoisEastern KentuckyEastern MichiganEastern WashingtonEdward WatersElizabeth City StateElonFayetteville StateFerrumFloridaFlorida A&MFlorida AtlanticFlorida InternationalFlorida StateFordhamFresno StateFurmanGardner-WebbGeorgetownGeorgiaGeorgia SouthernGeorgia StateGeorgia TechGramblingGreensboro CollegeHamptonHanover CollegeHarvardHawai'iHoly CrossHoustonHouston ChristianHowardIdahoIdaho StateIllinoisIllinois StateIncarnate WordIndianaIndiana StateIowaIowa StateJackson StateJacksonville StateJames MadisonKansasKansas StateKennesaw StateKent StateKentuckyKentucky ChristianKentucky StateLSULafayetteLamarLane CollegeLangston UniversityLehighLibertyLincoln (CA)Lincoln (PA)LindenwoodLittle RockLong Island UniversityLouisianaLouisiana CollegeLouisiana TechLouisvilleMaineMaristMarshallMarylandMassachusettsMcKendreeMcNeeseMemphisMercerMercyhurstMerrimackMiamiMiami (OH)MichiganMichigan StateMiddle TennesseeMiles CollegeMinnesotaMississippi StateMississippi Valley StateMissouriMissouri StateMonmouthMontanaMontana StateMorehead StateMorehouse CollegeMorgan StateMurray StateNC StateNavyNebraskaNevadaNew HampshireNew HavenNew MexicoNew Mexico StateNichollsNorfolk StateNorth AlabamaNorth CarolinaNorth Carolina A&TNorth Carolina CentralNorth DakotaNorth Dakota StateNorth TexasNortheastern StateNorthern ArizonaNorthern ColoradoNorthern IllinoisNorthern IowaNorthern MichiganNorthwesternNorthwestern StateNotre DameOhioOhio StateOklahomaOklahoma StateOld DominionOle MissOregonOregon StatePenn StatePennsylvaniaPittsburghPortland StatePrairie View A&MPresbyterianPrincetonPurdueRhode IslandRiceRichmondRobert MorrisRutgersSE LouisianaSMUSacramento StateSacred HeartSaginaw Valley StateSam HoustonSamfordSan DiegoSan Diego StateSan José StateSouth AlabamaSouth CarolinaSouth Carolina StateSouth DakotaSouth Dakota StateSouth FloridaSoutheast Missouri StateSouthernSouthern IllinoisSouthern MissSouthern UtahSt. AnselmSt. Francis (PA)St. Thomas (MN)StanfordStephen F. AustinStetsonStonehillStony BrookSul Ross StateSyracuseTCUTarleton StateTempleTennesseeTennessee StateTennessee TechTexasTexas A&MTexas SouthernTexas StateTexas TechTexas WesleyanThe CitadelThomas More CollegeToledoTowsonTroyTruman StateTulaneTulsaTuskegeeUABUAlbanyUC DavisUCFUCLAUConnUL MonroeUNLVUSCUT MartinUT Rio Grande ValleyUTEPUTSAUpper Iowa UniversityUtahUtah StateUtah TechVMIValparaisoVanderbiltVillanovaVirginiaVirginia StVirginia TechVirginia University Of LynchburgWagnerWake ForestWarner UniversityWashingtonWashington StateWebber InternationalWeber StateWest GeorgiaWest LibertyWest VirginiaWestern CarolinaWestern IllinoisWestern KentuckyWestern MichiganWestern OregonWilliam & MaryWisconsinWoffordWyomingYaleYoungstown State
Every team in the fit is a node; every game played through week 4 is an edge. Teams in different components have no chain of results connecting them at all. Not through a common opponent, not through anyone’s common opponent, not at any length. The graph is welded into one component, all 310 teams mutually comparable. That is the condition under which a ranking is a measurement rather than an extrapolation. Watching this knit together, week by week, is the content.

Diagnostics

teams in the fitEvery team with at least one game against an FBS or FCS opponent.310
games played513
connected componentsSeparate islands of the schedule graph. Two teams in different components have no chain of results connecting them at all, at any length.1
largest componentShare of the field that is mutually comparable through played games.310 teams (100.0%)
bridge gamesSingle games whose removal would split the graph in two, each holding at least two teams on. Every rating on the far side rests on that one result.1
fitted λ₂ (results core)Chosen by cross-validation every week. λ is a ratio of variances, a statement about how much the model does not know, containing no team-specific information whatsoever. It is large when the data is thin and falls as the season accumulates. Regularization stays a statement about variance and never becomes a reputation prior.1.000
median 90% rank-interval widthOut of 136 ranked teams, from 1,000 replays of this exact schedule. This is the number that says whether the season has settled.113 places
teams whose interval spans the league90% interval at least 90% as wide as the whole field.20
teams no closer than three hops to the top tenThey have not played a top-ten team and share no opponent with one. Their position relative to the top of the poll is an extrapolation.190

These are the measures for the weeks before the poll opens, published whether or not they are flattering. The one worth understanding is the penalty the model puts on itself: chosen fresh every week by testing, identical for every team, and carrying nothing about any particular one. Pulling an unknown team toward the middle is a statement about the model’s own ignorance, which is why it can produce a number in week 2 without being told which conferences are supposed to be good.

Bridge games

A bridge is a game whose removal would split the graph in two. Every rating on the far side of one rests on that single result, rather than on a body of evidence, on one Saturday afternoon. There is 1 of them holding at least two teams on right now.

  • Stetson at Chattanooga (week 3). Chattanooga against Stetson is the only game linking 4 teams to the other 306. Undo that one result and the graph splits in two.

What would have to be true

The graph is welded: every team is connected to every other through some chain of results. That is the condition under which a ranking is a measurement rather than an extrapolation, and it is why the opening week is week 5 and not week 1.

Chattanooga against Stetson is the only game linking 4 teams to the other 306. Undo that one result and the graph splits in two.

Every one of these is worth more to this poll than its television slot suggests. They are the games that turn two separate islands of results into one comparable field, and no ranking that exists today will tell you which they are.

The provisional table

The estimator runs from week 1. The headline poll does not. What is behind this click is real output from the real model on the real schedule, and it is not the poll. It is labelled provisional on every row, on purpose, because the difference between computing something and being willing to publish it as a ranking is the entire point of this page.

Show the provisional table: 136 teams, ranked by schedule oddsPROVISIONAL. THE POLL OPENS IN WEEK 5. SEE docs/constraints.md.
ProvisionalPROVISIONAL. THE POLL OPENS IN WEEK 5. SEE docs/constraints.md.
#teamunderline: 90% rank interval · league median width 113how unlikely1 in this many
1MemphisAmerican Athletic4-0·Power 10.33 (14)·Gap +49.67·▼24Ranked 1. The model replayed this schedule 1,000 times, and this team finished between 4 and 113 of 136 in 90% of them.0.9341 in 9
2OklahomaSEC4-0·Power 4.01 (45)·Gap +55.99·▼2Ranked 2. The model replayed this schedule 1,000 times, and this team finished between 4 and 110 of 136 in 90% of them.0.8631 in 7
3Texas TechBig 124-0·Power 23.90 (1)·Gap +36.10Ranked 3. The model replayed this schedule 1,000 times, and this team finished between 1 and 97 of 136 in 90% of them.0.7741 in 6
4GeorgiaSEC3-0·Power 4.03 (44)·Gap +55.97·▼3Ranked 4. The model replayed this schedule 1,000 times, and this team finished between 4 and 129 of 136 in 90% of them.0.7661 in 6
5Mississippi StateSEC4-0·Power 9.14 (18)·Gap +50.86·▼17Ranked 5. The model replayed this schedule 1,000 times, and this team finished between 6 and 118 of 136 in 90% of them.0.7461 in 6
6Iowa StateBig 124-0·Power 2.66 (54)·Gap +57.34·▲1Ranked 6. The model replayed this schedule 1,000 times, and this team finished between 10 and 121 of 136 in 90% of them.0.7351 in 5
7VanderbiltSEC4-0·Power 21.04 (2)·Gap +38.96·▼11Ranked 7. The model replayed this schedule 1,000 times, and this team finished between 6 and 99 of 136 in 90% of them.0.7291 in 5
8UNLVMountain West4-0·Power -1.79 (96)·Gap +61.79·▼16Ranked 8. The model replayed this schedule 1,000 times, and this team finished between 22 and 130 of 136 in 90% of them.0.6961 in 5
9Ole MissSEC4-0·Power 0.38 (79)·Gap +59.62·▼2Ranked 9. The model replayed this schedule 1,000 times, and this team finished between 8 and 124 of 136 in 90% of them.0.6931 in 5
10North TexasAmerican Athletic4-0·Power 12.43 (10)·Gap +47.57·▲4Ranked 10. The model replayed this schedule 1,000 times, and this team finished between 5 and 102 of 136 in 90% of them.0.6911 in 5
11LSUSEC4-0·Power -0.30 (83)·Gap +60.30·▲3Ranked 11. The model replayed this schedule 1,000 times, and this team finished between 12 and 125 of 136 in 90% of them.0.6841 in 5
12HoustonBig 123-0·Power 2.52 (55)·Gap +57.48·▼17Ranked 12. The model replayed this schedule 1,000 times, and this team finished between 9 and 130 of 136 in 90% of them.0.6351 in 4
13Georgia TechACC4-0·Power 0.51 (78)·Gap +59.49·▼7Ranked 13. The model replayed this schedule 1,000 times, and this team finished between 10 and 127 of 136 in 90% of them.0.6141 in 4
14MarylandBig Ten4-0·Power 8.31 (22)·Gap +51.69·▼20Ranked 14. The model replayed this schedule 1,000 times, and this team finished between 12 and 120 of 136 in 90% of them.0.6121 in 4
15Texas A&MSEC3-0·Power 3.89 (46)·Gap +56.11·▲14Ranked 15. The model replayed this schedule 1,000 times, and this team finished between 7 and 129 of 136 in 90% of them.0.6021 in 4
16USCBig Ten4-0·Power 14.66 (6)·Gap +45.34·▼10Ranked 16. The model replayed this schedule 1,000 times, and this team finished between 9 and 97 of 136 in 90% of them.0.6001 in 4
17BYUBig 123-0·Power 13.78 (7)·Gap +46.22·▼2Ranked 17. The model replayed this schedule 1,000 times, and this team finished between 5 and 120 of 136 in 90% of them.0.5611 in 4
18MiamiACC4-0·Power 6.76 (28)·Gap +53.24·▲16Ranked 18. The model replayed this schedule 1,000 times, and this team finished between 13 and 120 of 136 in 90% of them.0.5541 in 4
19IndianaBig Ten4-0·Power 8.45 (21)·Gap +51.55·▲9Ranked 19. The model replayed this schedule 1,000 times, and this team finished between 7 and 107 of 136 in 90% of them.0.5511 in 4
20TCUBig 123-0·Power 5.62 (36)·Gap +54.38·▲7Ranked 20. The model replayed this schedule 1,000 times, and this team finished between 10 and 111 of 136 in 90% of them.0.5201 in 3
21OregonBig Ten4-0·Power 16.33 (3)·Gap +43.67·▲12Ranked 21. The model replayed this schedule 1,000 times, and this team finished between 14 and 111 of 136 in 90% of them.0.5201 in 3
22NavyAmerican Athletic3-0·Power 8.72 (20)·Gap +51.28·▼21Ranked 22. The model replayed this schedule 1,000 times, and this team finished between 11 and 116 of 136 in 90% of them.0.4741 in 3
23Florida StateACC3-0·Power 13.74 (8)·Gap +46.26·▲6Ranked 23. The model replayed this schedule 1,000 times, and this team finished between 11 and 121 of 136 in 90% of them.0.4531 in 3
24UtahBig 123-1·Power 11.23 (13)·Gap +9.81·▼11Ranked 24. The model replayed this schedule 1,000 times, and this team finished between 2 and 106 of 136 in 90% of them.0.4501 in 3
25MissouriSEC4-0·Power 3.17 (48)·Gap +56.83·▼2Ranked 25. The model replayed this schedule 1,000 times, and this team finished between 18 and 124 of 136 in 90% of them.0.4391 in 3

Top 25 of 136. The full table, the résumé column, the schedule receipts and the sort controls arrive with the week 5 launch.

How this works, and where it is weak →

run 52073e40 · published 2026-08-19 23:11:26 UTC · code e215160 · config 1ef7cf23…
q_ref 7.35 (Nebraska) · β_w 7 · C 32 · h 8.019 · σ 21.800 · λ₁ 175 · λ₂ 1.0 · k 93.20 · w₁ 0.8512 · w₂ -0.4721
reproduce: uv run cfbpoll rank --season 2025 --through-week 4 --out out/