The AHSAA girls flag football playoffs start on November 3 and finish with four state championship games on December 1. Three public-school brackets (6A, 5A and 1A-4A) and one private-school bracket (A-AA) decide the titles. Below each bracket is a table of how often each team should reach each round if the ratings are right. Pick a classification to switch brackets.
Loading brackets and odds…
How to read the odds
Each number is the probability that a team is still playing after that round. A team with a first-round bye shows 100% for the first round. The last column is the chance of winning the state title. A team listed at 30% should win the title about three times in ten, which also means it should fail to about seven times in ten. That is why a forecast is a range of outcomes and not a pick. The playoff simulation explainer covers why even strong favorites lose brackets more often than people expect.
Once a playoff game is final in the AHSAA scores feed, its winner is fixed and that team's odds for the round become 100%. The other teams' odds are recalculated from there. Brackets on BracketNest are updated by hand as results come in, so they can trail the odds table by a day.
Why the Power Rating
The rankings page publishes four ratings for every team: the Power Rating, the Power Rating with no carryover from last season (cold start), an Elo rating, and Elo with no carryover. These odds use the Power Rating, the points-scale model that fits every game statewide at once with strength of schedule built in, and starts each team from a discounted version of last season's rating. The choice came from testing the four ratings on games they had not seen.
The test is a walk-forward backtest on the 2026 season. For each week, each rating was rebuilt from only the games played before that week and then asked to predict that week's games. Lower Brier score is better. It measures how far the stated probabilities were from what happened, so it rewards being well calibrated and not just picking the right winner.
| Rating | Winners picked, whole season | Brier, whole season | Winners picked, Oct 1 on | Brier, Oct 1 on |
|---|---|---|---|---|
| Power Rating | 79.6% | 0.140 | 86.0% | 0.099 |
| Power Rating (cold start) | 75.4% | 0.161 | 83.7% | 0.105 |
| Elo | 77.9% | 0.150 | 83.0% | 0.120 |
| Elo (cold start) | 73.6% | 0.174 | 81.3% | 0.136 |
The whole-season columns cover 769 decided games from the second week of the season through October 27. The October columns cover the 300 games closest to playoff conditions. The Power Rating was best on both measures in both windows. Last season's playoffs point the same way: rebuilt from regular-season games only, the cold-start Power Rating picked 52 of the 60 decided 2025 playoff games (86.7%, Brier 0.095), against 49 for cold-start Elo (81.7%, Brier 0.139). The seeded versions could not be tested on 2025, because there is no 2024 season to seed from.
There is also a structural reason to prefer it in November. Elo moves a rating a little after each game and never goes back, so an early win over a team that later turned out to be weak still counts as a win over a strong team. The Power Rating refits every game together, so a team's rating reflects its whole schedule, including how its opponents fared afterwards. By the end of the regular season that full-schedule view is worth more than Elo's emphasis on recent form.
Data and methods
Brackets. The pairings come from the AHSAA brackets as published by alpreps.com on October 28, 2026 (5A, 1A-4A, A-AA). Labels such as R4-1 mean the Region 4 champion. In 6A and A-AA the region champions and runners-up have first-round byes. Every team's region was checked against the AHSAA flag football region alignment, and every first-round pairing was checked against the playoff games scheduled in the AHSAA scores feed. All 44 first-round games matched.
Win probability. For two teams A and B, the Power Rating gap is the predicted margin. The chance that A wins is the normal probability that the real margin lands above zero, using the model's residual spread (about 9.7 points this season). In the first round the host is known from the AHSAA schedule, so the host gets the model's home-field edge (about 1.4 points). Later-round hosts depend on travel and on who advances, so those games are treated as neutral.
Bracket odds. The odds are calculated exactly, not by drawing random brackets. The chance that a team wins a round is the chance it won its own part of the bracket, multiplied by its chance of beating each possible opponent from the other part, weighted by how likely that opponent is to be there. This gives the same answer a Monte Carlo simulation approaches with infinitely many runs, without its sampling noise. As a check, 100,000 simulated brackets per classification agreed with the exact figures to within 0.4 percentage points.
What the odds leave out. The ratings treat a team as one number. They know nothing about injuries, weather, or a team that is playing better than its results. Teams also file their own scores, and a missing or late score can move a rating. Read a percentage as a summary of the evidence in the scores, not a statement about the players. The rankings methodology describes the rating model in full, and the editorial policy covers corrections.