How the ALPreps Index Works

RPI is a separate, simpler measure kept alongside the index: 25% your win percentage, 50% your opponents’ win percentage, and 25% your opponents’ opponents’ win percentage.

Out-of-state opponents count toward your own record but are excluded from the opponent-strength terms, since there is no in-system record for them.

Each scheduled game shows a projected margin from the rating gap plus home-field advantage. Once a result is entered, it is labeled against that projection — dominant, exceeded, as expected, or below expectation. Projections never feed back into the ratings.

The odds are the rest of the season played ten thousand times. Each game still to come is decided by a coin weighted by the two ratings and home field, the region tables are rebuilt, and the brackets are seeded and played out. A team’s odds are the share of those seasons in which the thing happened.

The one number behind it is how often a rating gap turns into a win. It is fitted against the 2025 season rather than guessed, by rating the weeks before each week and predicting that week cold. It picks about four winners in five, and what it calls a 70% chance happened about 70% of the time.

Teams level in a region are separated by the AHSAA’s published procedure, (a) through (q) — head-to-head, then records against each ranked team in the region, then non-region common opponents, then the strength of the teams they beat. The final step is a coin flip administered by the association; the Index rating stands in for it here, so a tie does not reshuffle itself every time the board is published.

How many teams each region sends differs by classification: 6A sends six, Class AA sends every team it has — so its bracket is about seeding rather than qualifying — and the rest send four.

“Clinched” and “eliminated” are not read off the simulation. Ten thousand seasons without an outcome is not a proof, and those two words deserve one, so both come from what the remaining games make arithmetically possible.

The percentages on the Playoff Odds board are arithmetic: the rest of the season played ten thousand times, and the share of those seasons in which a thing happened. Nobody’s opinion is in them.

The coloured status on the Bracketology page is the opposite — a judgement about where a team will finish. It is not a ranking of how good a team is. In a classification where every team reaches the bracket, a side certain to finish last is “High” for exactly the same reason as the side certain to finish first: there is high confidence about where it lands. Three teams who could finish in any order among themselves are all “Medium”, even when all three are going through.

“Clinched” there means a team has locked up that particular place — not that it has reached the playoffs. The two can be far apart, and in a class where everyone qualifies the second is true from the opening whistle while the first may not be settled until the last night.

The association sometimes rules a program out of championship play for the season. That does more than keep it out of the bracket: its region schedule is void for everyone. The barred team finishes 0-0 in region, and the teams that played it take an overall win or loss and no region result at all. For the tie-breaking procedure those games are not there.

The rating does not move. The football was played, and it is still the best evidence of how good a team is — so the Index reads those games exactly as it reads any other. This is the same split as a forfeit, asked of a different question: not who won, but whether the game counted toward the bracket.

A forfeit is a ruling about the record, not about the football. A team that wins on the field and gives the game up afterwards — almost always an eligibility finding weeks later — keeps the scoreline and loses the win.

So the two are separated. Records, standings, region order, RPI and Résumé all count the forfeit, because those are statements about what a team officially holds. The Index does not: it measures how well a team plays, and a paperwork ruling does not change how a game was played. A record marked with an asterisk includes a game decided this way.