2026 season · Through Week 3
Blogger vs. the Math
My rankings are opinion. The math is an Elo rating for every team, built from last season, every final score and the box scores, then played forward 10,000 times. Here is where we disagree, who has been right, what happens next, and who is likely to make the playoffs.
You are looking at how things stood after Week 3, using only what was known then. Back to the latest.
The Raiders
I have them No. 5. The math has them No. 14.
That puts me 9 spots higher than the math.
Make the playoffs
70%
Results alone: 80%
Win the AFC West
35%
Results alone: 44%
Projected wins
9.9
Now 3–0, 14 to play
Schedule left
No. 2
of 32, 1 is hardest
Where they finish: chance of each win total
Bright bars are totals still possible. Numbers are percent.
Why two numbers? The full model starts from last season’s ratings and looks at yards per play, not just the score, so it is slower to believe a hot start. Results alone put the Raiders at 80% to make the playoffs; the full model says 70%. Neither knows about injuries.
Next up: who wins?
The math’s forecast for the coming games, next to my pick. Where we pick different teams, you get a vote: who do you trust?
Week 4
- PITSteelersatCLEBrowns
Thu, Oct 1, 5:15 PM PT
PIT 45%CLE by 255% CLEWe splitMe: PIT · Math: CLE - INDColtsatWASCommanders
Sun, Oct 4, 6:30 AM PT
IND 44%WAS by 256% WASWe agreeMe: WAS · Math: WAS - ARICardinalsatNYGGiants
Sun, Oct 4, 10:00 AM PT
ARI 42%NYG by 358% NYGWe agreeMe: NYG · Math: NYG - DALCowboysatHOUTexans
Sun, Oct 4, 10:00 AM PT
DAL 45%HOU by 255% HOUWe splitMe: DAL · Math: HOU - GBPackersatTBBuccaneers
Sun, Oct 4, 10:00 AM PT
GB 46%TB by 1.554% TBWe splitMe: GB · Math: TB - JAXJaguarsatCINBengals
Sun, Oct 4, 10:00 AM PT
JAX 45%CIN by 255% CINWe agreeMe: CIN · Math: CIN - LARRamsatPHIEagles
Sun, Oct 4, 10:00 AM PT
LAR 45%PHI by 1.555% PHIWe splitMe: LAR · Math: PHI - NEPatriotsatBUFBills
Sun, Oct 4, 10:00 AM PT
NE 42%BUF by 358% BUFWe agreeMe: BUF · Math: BUF - NYJJetsatCHIBears
Sun, Oct 4, 10:00 AM PT
NYJ 42%CHI by 2.558% CHIWe agreeMe: CHI · Math: CHI - TENTitansatBALRavens
Sun, Oct 4, 10:00 AM PT
TEN 38%BAL by 462% BALWe agreeMe: BAL · Math: BAL - MIADolphinsatMINVikings
Sun, Oct 4, 1:05 PM PT
MIA 39%MIN by 3.561% MINWe agreeMe: MIN · Math: MIN - DENBroncosatSF49ers
Sun, Oct 4, 1:25 PM PT
DEN 39%SF by 3.561% SFWe agreeMe: SF · Math: SF - KCChiefsatLVRaiders
Sun, Oct 4, 1:25 PM PT
KC 47%LV by 153% LVWe splitMe: KC · Math: LV - LACChargersatSEASeahawks
Sun, Oct 4, 1:25 PM PT
LAC 37%SEA by 4.563% SEAWe agreeMe: SEA · Math: SEA - DETLionsatCARPanthers
Sun, Oct 4, 5:20 PM PT
DET 44%CAR by 256% CARWe splitMe: DET · Math: CAR - ATLFalconsatNOSaints
Mon, Oct 5, 5:15 PM PT
ATL 45%NO by 1.555% NOWe agreeMe: NO · Math: NO
Percentages are the math’s chance each side wins, with home field, rest and travel included. The line is the margin the ratings expect, to the nearest half point. My pick is the higher team on my board.
Who has been right?
The rankings are only useful if they predict something. So every week’s board is graded on the next week’s games.
Me
8–8
50% of winners picked
The math
6–10
38% of winners picked
When we disagree
2–0
2 games, my pick first
I'm ahead by 2 games. Each week’s rankings are graded on the next week’s games, 16 so far. A pick is simply the higher-ranked team: no home field, no point spread, ties skipped.
The 2 games we disagreed on
- Wk 3NYJ 24, DET 31I picked DET, the math picked NYJ.I was right
- Wk 3LV 35, NO 27I picked LV, the math picked NO.I was right
Playoff odds
The rest of the season, played 10,000 times from the ratings. Each division is sorted by its chance to win it; the big number is the chance to make the playoffs.
AFC
| Team | Record | Projected wins | Playoff chance | Division chance |
|---|---|---|---|---|
| BUFBills | 3–0 | 10.3 W | 79% | Division 65% |
| NEPatriots | 1–2 | 8.1 W | 36% | Division 17% |
| NYJJets | 1–2 | 7.8 W | 31% | Division 14% |
| MIADolphins | 0–3 | 6.5 W | 9% | Division 4% |
| Team | Record | Projected wins | Playoff chance | Division chance |
|---|---|---|---|---|
| BALRavens | 2–1 | 9.5 W | 64% | Division 34% |
| CINBengals | 2–1 | 9.1 W | 55% | Division 26% |
| CLEBrowns | 2–1 | 9.0 W | 47% | Division 20% |
| PITSteelers | 2–1 | 8.9 W | 45% | Division 19% |
| Team | Record | Projected wins | Playoff chance | Division chance |
|---|---|---|---|---|
| JAXJaguars | 2–1 | 9.2 W | 70% | Division 62% |
| INDColts | 1–2 | 7.6 W | 27% | Division 18% |
| HOUTexans | 0–3 | 6.8 W | 18% | Division 11% |
| TENTitans | 0–3 | 6.6 W | 14% | Division 9% |
| Team | Record | Projected wins | Playoff chance | Division chance |
|---|---|---|---|---|
| KCChiefs | 3–0 | 10.3 W | 79% | Division 48% |
| LVRaiders | 3–0 | 9.9 W | 70% | Division 35% |
| DENBroncos | 2–1 | 8.9 W | 46% | Division 16% |
| LACChargers | 0–3 | 6.4 W | 9% | Division 2% |
NFC
| Team | Record | Projected wins | Playoff chance | Division chance |
|---|---|---|---|---|
| PHIEagles | 2–1 | 9.2 W | 54% | Division 38% |
| NYGGiants | 2–1 | 8.8 W | 44% | Division 29% |
| DALCowboys | 1–2 | 7.9 W | 31% | Division 18% |
| WASCommanders | 1–2 | 7.8 W | 26% | Division 15% |
| Team | Record | Projected wins | Playoff chance | Division chance |
|---|---|---|---|---|
| MINVikings | 3–0 | 10.3 W | 76% | Division 47% |
| CHIBears | 2–1 | 9.1 W | 57% | Division 25% |
| DETLions | 2–1 | 9.2 W | 51% | Division 21% |
| GBPackers | 1–2 | 7.9 W | 24% | Division 7% |
| Team | Record | Projected wins | Playoff chance | Division chance |
|---|---|---|---|---|
| CARPanthers | 1–2 | 8.1 W | 41% | Division 31% |
| NOSaints | 1–2 | 8.1 W | 39% | Division 30% |
| ATLFalcons | 1–2 | 8.0 W | 35% | Division 27% |
| TBBuccaneers | 0–3 | 6.7 W | 17% | Division 12% |
| Team | Record | Projected wins | Playoff chance | Division chance |
|---|---|---|---|---|
| SF49ers | 3–0 | 10.7 W | 84% | Division 56% |
| SEASeahawks | 2–1 | 9.8 W | 69% | Division 31% |
| LARRams | 1–2 | 8.2 W | 36% | Division 9% |
| ARICardinals | 1–2 | 7.4 W | 18% | Division 4% |
Compare any two teams
Pick two teams for their ratings, odds and what the math says if they played each other.
On a neutral field the math gives the Chiefs a 53% chance.
At Raiders home: 52% Raiders. At Chiefs home: 41% Raiders. Expected margin on a neutral field: Chiefs by 1.1. They meet in Week 4 in Raiders country: 53% for the Raiders.
| LV | Stat | KC |
|---|---|---|
| No. 5 | My rank | No. 3 (better) |
| No. 14 | The math’s rank | No. 4 (better) |
| 1503 | Math rating | 1532 (better) |
| 3–0 | Record | 3–0 |
| 70% | Make the playoffs | 79% (better) |
| 35% | Win the division | 48% (better) |
| 9.9 | Projected wins | 10.3 (better) |
| No. 2 | Schedule left (1 = hardest) | No. 5 |
| +11.3 | Scoring margin per game | +12.7 (better) |
| −1.7 | Margin by yards per play | +13.6 (better) |
| +1.7 (better) | Turnover margin per game | +1.0 |
Bold is the better number. Margins are points per game; the yards-per-play margin is net yards per play converted to the scoring scale.
Where we disagree most
Teams where my rank and the math’s rank are 6 or more spots apart. Each one has a card you can share.
- NEI'm lower
The math loves the Patriots more than I do.
I have them No. 26. The math has them No. 8, 18 spots higher.
- WASI'm higher
The math doesn't believe in the Commanders.
I have them No. 17. The math has them No. 28, 11 spots lower.
- CINI'm higher
The math doesn't believe in the Bengals.
I have them No. 6. The math has them No. 16, 10 spots lower.
- LVI'm higher
The math doesn't believe in the Raiders.
I have them No. 5. The math has them No. 14, 9 spots lower.
- INDI'm higher
The math doesn't believe in the Colts.
I have them No. 20. The math has them No. 29, 9 spots lower.
- ARII'm higher
The math doesn't believe in the Cardinals.
I have them No. 23. The math has them No. 32, 9 spots lower.
- ATLI'm lower
The math loves the Falcons more than I do.
I have them No. 25. The math has them No. 17, 8 spots higher.
- BALI'm lower
The math loves the Ravens more than I do.
I have them No. 12. The math has them No. 5, 7 spots higher.
- SEAI'm lower
The math loves the Seahawks more than I do.
I have them No. 8. The math has them No. 2, 6 spots higher.
- NOI'm higher
The math doesn't believe in the Saints.
I have them No. 15. The math has them No. 21, 6 spots lower.
- LACI'm lower
The math loves the Chargers more than I do.
I have them No. 30. The math has them No. 24, 6 spots higher.
All 32 teams
Open any team for its rank history, every result and what each one did to its rating, what the box scores say, and the games still to come.
Gap: + means I rank them higher than the mathTap a team for its story.
| Me | Team | Where each ranks | Math | Gap |
|---|---|---|---|---|
| 1 | 3 | I'm 2 higher | ||
| 2 | 1 | I'm 1 lower | ||
| 3 | 4 | I'm 1 higher | ||
| 4 | 7 | I'm 3 higher | ||
| 5 | 14 | I'm 9 higher | ||
| 6 | 16 | I'm 10 higher | ||
| 7 | 10 | I'm 3 higher | ||
| 8 | 2 | I'm 6 lower | ||
| 9 | 11 | I'm 2 higher | ||
| 10 | 12 | I'm 2 higher | ||
| 11 | 6 | I'm 5 lower | ||
| 12 | 5 | I'm 7 lower | ||
| 13 | 9 | I'm 4 lower | ||
| 14 | 15 | I'm 1 higher | ||
| 15 | 21 | I'm 6 higher | ||
| 16 | 18 | I'm 2 higher | ||
| 17 | 28 | I'm 11 higher | ||
| 18 | 13 | I'm 5 lower | ||
| 19 | 19 | Same | ||
| 20 | 29 | I'm 9 higher | ||
| 21 | 26 | I'm 5 higher | ||
| 22 | 23 | I'm 1 higher | ||
| 23 | 32 | I'm 9 higher | ||
| 24 | 20 | I'm 4 lower | ||
| 25 | 17 | I'm 8 lower | ||
| 26 | 8 | I'm 18 lower | ||
| 27 | 22 | I'm 5 lower | ||
| 28 | 25 | I'm 3 lower | ||
| 29 | 27 | I'm 2 lower | ||
| 30 | 24 | I'm 6 lower | ||
| 31 | 30 | I'm 1 lower | ||
| 32 | 31 | I'm 1 lower |
How the math works
Every team has an Elo rating, and the winner of a game takes points from the loser. Beat a better team, or win by more, and you take more. Home teams get about two points of credit, so a road win counts for a little extra. The margin of victory is dampened, so a blowout of a bad team doesn’t run the ratings away.
The full model adds three things to that. It starts the season from last year’s final ratings, pulled almost all the way back toward average (only a tenth of last year’s edge is kept), so this season counts for far more than last, since rosters turn over and results are partly luck. It nudges each rating toward what the box scores say: a team that wins ugly, with fewer yards per play than its score suggests, is rated a little lower than its record, and one that outgains everyone but loses close games a little higher (48 box scores so far). And it credits rest and travel: a bye week, a short week on Thursday night, the distance the visitor flew.
I tested the additions on last season’s games, predicting each from the games before it. A gentle yards-per-play nudge (30% of the gap, and never more than 40 rating points) made the predictions a little better and a heavy one made them worse, so the weight is deliberately small. Rest and travel could not be told apart from noise, so they count at half strength. Turnovers are shown on each team’s panel but don’t move the rating: they decide games, but rarely repeat from team to team.
Playoff odds come from playing every game still on the schedule 10,000 times. Each result is drawn around the margin the ratings expect (about 25 rating points per point on the scoreboard, with the usual spread of NFL results), and then the standings are worked out: division winners, then three wild cards, in each conference. Ratings are held fixed while a season plays out.
It is a simplification. The math knows nothing about injuries or quarterbacks, and ties in the standings are broken by win percentage and then point differential, not the league’s full head-to-head and division-record rules. Early in the year it is mostly last season plus a few games, so expect it to wobble through the first month. It is a check on my opinions, not a replacement for them.
Results and box scores come from ESPN’s public scoreboard and update after every final. Votes are anonymous and one per browser; they are a reader poll, not part of the model.