Methodology · Formula MFPI-3.2

How MFPI is calculated

Week 9, 2026 · 225 rated teams · updated October 27, 2026

MFPI asks one question: based on whom a team played and how it performed, how strong is that team right now? Eighty percent of the answer comes from MHSAA classifications, schedules and scores. The remaining twenty percent comes from the MaxPreps Mississippi rankings' statewide rank and strength-of-schedule values.

The eight components

Each raw component becomes a statewide percentile among active MHSAA 1A–7A teams. Ties receive average rank, and missing early-season values are transparently neutral rather than punitive. The weighted sum is constrained to 1.0–100.0 and displayed to one decimal, while the ranking itself uses full precision.

Opponent-adjusted margin

From each team’s perspective, the margin in a game is compressed rather than taken at face value, then placed against the opponent’s current rating and adjusted for where the game was played. A road team receives a small bonus and a home team a small penalty; the recorded score never changes. Compressing the margin gives blowouts diminishing returns, so a 60-point win is worth more than a 30-point win but not twice as much.

Team ratings and opponent ratings are solved together, repeating until every team’s rating stops moving, then centred across the state. This is why beating a strong team by a little can be worth more than beating a weak team by a lot.

The enrollment-class starting assumption

Early in a season the statewide game graph is barely connected — teams have played two or three opponents, and those opponents have played almost nobody in common. Until that resolves, MFPI assumes larger enrollment classes are stronger on average, on a sliding scale from 7A down to 1A.

This is a fading prior, not a permanent bonus. It carries the weight of about two games through a team’s first two games, one game through five, and almost nothing after that. Real margins and opponent chains progressively replace it, and an excellent lower-class team can and does pass higher-class teams. Playing up or down follows directly: face a higher class and both your opponent-adjusted performance and your schedule strengthen.

Confirmed bye weeks

A team that does not play should not swing wildly on other teams’ results. On a confirmed bye with unchanged verified results, the published score is 90% of the previous week’s rating and 10% of the fresh recalculation. The adjustment is symmetric — it dampens gains and losses equally.

It protects the score, not the rank: other teams can still pass an idle team. A bye also has to be proven, not assumed. A team’s schedule must show games before and after the empty week; simply missing from a feed is not a bye, and neither is a cancellation or postponement. Teams with no previous games get no protection.

Data sources

• Teams, classifications and regions: the official MHSAA 2025–27 football classification list. Only active MHSAA Class 1A–7A programs are ranked; MAIS, out-of-state and other opponents appear on schedules but are never ranked.
• Schedules and scores: the official MHSAA score center. A game counts only when it is final with both scores. Jamborees and scrimmages never count.
• Media rank and media strength of schedule: the MaxPreps Mississippi football rankings, matched to the MHSAA list exactly once per team.
• Gaps: when the MHSAA score center lacks a final for a game that should be over, MaxPreps team schedules are checked. A result is accepted only for the same two teams on the scheduled date or one day either side, and conflicting reports are rejected. An absent score is never read as a cancellation.
• Out-of-state opponents: a published MaxPreps rating, when available, gives a capped starting estimate that fades as real results connect.

Missing data and provisional ratings

A missing score is never treated as 0–0 or as a loss, and postponed or cancelled games do not count. A component with no data for a team (for example scoring averages before its first verified game) sits at a neutral 50th percentile rather than a punitive zero. When fewer than 95% of the games due by the cutoff have verified scores, the whole week is labelled provisional. A team with no verified results is shown as “results unavailable”, its averages are shown as —, and its rating is labelled provisional because it rests on class, media and neutral inputs rather than games.

How changes and dates are shown

Rank movement is last week's rank minus this week's. Rating change is this week's displayed one-decimal rating minus last week's displayed rating, so the numbers on a page always add up. They are reported separately: a team can hold its rank while its rating rises. Both compare against the latest audited publication of the previous week; a team with no comparable previous row is shown as NEW. Game dates are calendar dates in Central time. Opponent ranks shown beside past games are their current ranks, not their ranks on game day.

Ranking, ties and corrections

State and class lists use the same MFPI score — a class rating is never recalculated separately. Ties break on unrounded rating, then opponent-adjusted performance, then strength of schedule, then head-to-head where applicable, then capped scoring margin, and finally team name so output is deterministic. Weekly movement compares against the previous published snapshot; a team absent from it shows as NEW.

When a score is corrected after publication, the week is reissued as an audited correction run rather than edited in place. Each snapshot records its own formula version, status and coverage — this week verified 980 of 1002 expected games.

See the full rankings, browse every team, or read more about the project.