Playoff Picture
Twelve spots. 50,000 simulated 2026 seasons, rerun after every week: each game already played is held at its actual result and the rest of the schedule is played out, through conference championships and a fitted selection committee.
A percentage here is the share of those seasons that end with the team in the field — not a ranking, and not a claim about any single game. Through week 2, 185 games played. Rebuilt 2026-09-13.
The field
Grouped by how often each team makes it, not by how many losses it can absorb — those are different questions. Nothing sits in the top band yet, and it should not: one week in, every team still has eleven games to play.
Near-lock
in 85%+ of simulated seasons1Faint
under 2%, but not zero20Every team
All 138 FBS teams the simulation ran, not a top twenty-five. Search yours and it answers with what it tells you about the favourites — how often 50,000 simulated seasons end with them in the field, what winning out would be worth, and the record those seasons project. Open the team for its full loss curve.
138 of 138 teams. If they win out is how often the team makes the field in the simulated seasons where they lose no more games — early in a year that is 100% for nearly every power-conference team, and it is the number that starts separating them. Affords is the most losses they can finish with and still be better than even.
Open a team for its full loss curve, its odds week by week, and what the model made of its last game. A wider screen adds the projected record and the columns for conference title and first-round bye.
Key games
The turning points of Navy’s season, measured as the gap between the two outcomes: how much playoff probability moves depending on who wins. A game Navy is not playing in can still move them, and later in the year those are the ones nobody works out for themselves.
Navy’s own turning points
Points of playoff probability between winning and losing.
- wk7at UTSA2.4
- wk12vs Memphis2.0
- wk15at Army1.9
- wk8vs North Texas1.9
- wk6vs Tulsa1.8
- wk5at Air Force1.7
Who to root for
Games Navy is not in that move its odds most. Same units, much smaller scale.
- wk4Illinois over Ohio State0.8
- wk9Texas State over Boise State0.6
- wk3Oklahoma over New Mexico0.5
- wk11UNLV over New Mexico0.4
- wk5UTEP over New Mexico0.4
- wk3San Diego State over James Madison0.4
The biggest swings in the sport
One row per game, naming the team it moves most among the teams not playing in it. This is the list that says which weekend to clear, whoever you support.
- wk3New Mexico at Oklahomamoves Boise State7.4
- wk4Northwestern at Indianamoves Ohio State5.2
- wk7Florida State at Miamimoves SMU4.4
- wk3Miami at Wake Forestmoves SMU4.3
- wk6LSU at Kentuckymoves Alabama4.3
- wk3James Madison at San Diego Statemoves Boise State4.3
Win out and it still might not be enough
For a power-conference team, running the table settles it: 44 of the 45 make the bracket in every simulated season they finish unbeaten. Outside those leagues it settles nothing, because the Group of 6 gets one bid and winning out only enters you into the argument for it.
- Liberty0-1 · 2% now24%
- Louisiana0-1 · 2% now36%
- Western Michigan0-1 · 6% now54%
- Memphis2-1 · 3% now55%
- North Texas1-1 · 2% now62%
- Army0-1 · 1% now65%
- Boise State1-1 · 20% now80%
- San Diego State0-1 · 3% now83%
- App State1-0 · 2% now85%
- Navy0-1 · 1% now87%
The last column is how often they make it in the simulated seasons where they lose no more games.
The Group of 6 race
One bid, to the highest-ranked champion outside the four power leagues — so these teams are competing against each other on a different rule, not against the SEC’s eighth best. All 62 with a path.
Who moved
Change in playoff probability from Week 1 to Week 2, in percentage points.
The second line is how each team did against the margin the ratings implied, which is what the model actually updates on. Winning is not the same as gaining: 17 of the 44 teams that came through week 2 unbeaten came out with worse odds than they went in with.
Gained the most
- No game the model counts — its opponent was FCS
- Won 59-7 over Texas Stateexp +27+24.7
- Won 52-16 over North Texasexp +40−3.8
- Won 48-10 over East Carolinaexp +22+15.7
- Won 13-12 over Western Michiganexp +25−24.4
- Won 50-0 over Missouri Stateexp +42+7.6
- No game the model counts — its opponent was FCS
- No game the model counts — its opponent was FCS
Lost the most
- Won 34-27 over Boise Stateexp +27−19.5
- Won 51-0 over UTEPexp +41+10.5
- Won 56-3 over Ball Stateexp +60−7.2
- No game the model counts — its opponent was FCS
- Won 24-10 over Washington Stateexp +23−8.5
- 2-0 on the week+15.4
- Won 45-0 over Marshallexp +27+18.2
- Won 40-0 over Northern Illinoisexp +34+5.8
Since preseason
The eight best odds on the board, week by week, on a shared 0–100% scale.
What a loss costs
The number every other tracker assumes is the same for everyone. It is not: a third loss leaves one of these teams comfortably in and another effectively out, and the difference is who they had the chance to beat. Affords is the most losses a team can take and still be better than even.
What the simulation also says
- 1.53 undefeated teams in an average simulated season.
- The 2026 rule change alters the field in 9.5% of seasons compared with the 2025 rules.
The committee is fitted, not consulted: it recovers about eleven of the real committee’s top twelve on a walk-forward test. How this is built →
