What the numbers actually mean: PRIME Football is built to describe team strength from real games, not recreate a poll. The core ratings account for opponent quality so the same statistical performance is treated differently against a strong opponent than against a weak one.
PRIME’s rating system separately attributes Offense, Defense, and Special Teams, each built from Expected Points (EP): every possession, kickoff, punt, field goal, and turnover is valued against how many points an average FBS team would expect to score from that same situation, then opponent-adjusted by solving the whole FBS schedule graph recursively. The possession-efficiency core used to fit both Offense and Defense excludes garbage-time drives; those two ratings also add opponent-adjusted Success Rate and Explosiveness using frozen weights, while Defense and Offense also add a small, opponent-adjusted turnover component (every giveaway or takeaway counts as exactly one event; return yards, return touchdowns and field position add nothing) and Defense adds separately attributed safeties. Special Teams remains a separately attributed expected-points unit. The methodology was selected from an 11-season (2014–2025) leakage-safe walk-forward test against the prior possession-only model: a statistically significant accuracy improvement, repeatable across 9 of 11 seasons. Net Rating is the overall team-strength metric; Off Rating, Def Rating, and ST Rating are its three sides, and Net Rating = Off Rating + Def Rating + ST Rating exactly. A team’s own rating is never blended with a preseason rating, recruiting input or conference-strength term. The one prior-season input is the opponent baseline: when the schedule graph is solved, each opponent’s strength counts as half its final rating from the previous season and half its current-season rating, at a flat weight all season (2014 and 2021 have no prior season). This stabilizes who a team really faced while this season’s sample is thin, and it was chosen by the same 2014–2025 walk-forward test: it lowered next-week margin error in every season that has a prior season.
PRIME’s opponent-adjusted performance system: Offense and Defense blend Expected-Points possession value with opponent-adjusted Success Rate and Explosiveness; Special Teams is a separately attributed opponent-adjusted Expected-Points unit.
The overall team-strength rating. Net Rating = Off Rating + Def Rating + ST Rating. Zero is approximately FBS average; higher is better.
Offensive composite of garbage-time-excluded, opponent-adjusted Expected-Points possession value, opponent-adjusted Success Rate and opponent-adjusted Explosiveness. Higher is better.
Defensive composite fit from the same garbage-time-excluded possession sample as Offense, plus a small opponent-adjusted takeaway-rate component (each takeaway is one event; return yards, return touchdowns and field position add nothing), separately attributed safeties, Success Rate prevention and Explosiveness prevention. Higher is better.
Special Teams composite: opponent-adjusted net Expected-Points value across kickoffs, punts, field goals and blocks, with a rules-based safety-attribution hierarchy for the rare kicking-play-adjacent safety. Higher is better.
Strength of Schedule: each opponent's PRIME Net Rating and game location are converted to the win probability for an exactly-average FBS team; those probabilities are averaged and converted back to a neutral-site PRIME difficulty. Higher is tougher.
Strength of Record: wins above what an exactly-average FBS team would be expected to earn against the same schedule. It answers a résumé question, separate from Net Rating's performance question.
Adjusted Score Matrix: an opponent-adjusted scoring-margin rating (constrained least squares), with each game's margin capped at 28 points before fitting so a blowout can't dominate a team's number. A résumé lens like SOR, not a second opinion on who's better -- see “Process vs. résumé” below. Shown on Advanced, not the main Ratings page.
Net Rating is the primary opponent-adjusted team-strength rating. It is not a final-score rating: the backbone is an Expected-Points ledger valuing every possession, kickoff, punt, field goal, and turnover against what an average FBS team would expect from that situation, solved recursively across the FBS opponent network, separately for Offense, Defense, and Special Teams. Offense and Defense then add opponent-adjusted Success Rate and Explosiveness using frozen weights validated on 2014–2025 walk-forward results; Special Teams contributes its separately attributed expected-points value. Net Rating equals Off Rating plus Def Rating plus ST Rating. ASM is intentionally different: it is an opponent-adjusted scoring-margin résumé lens with each game’s margin capped at 28 points.
Safety attribution. About 28% of safety plays cannot be text-classified as clearly defensive-caused or kicking-play-caused and fall back to a phase-based assignment, flagged low-confidence. This is a documented, published limitation, not a silent gap — see the rating snapshot’s own metadata.
Scoreboard reconciliation. A small residual (3.96% of games, in the validation audit) exists between the ledger’s play-by-play score reconstruction and the official final score, traced to a real source-data quirk (score fields occasionally jump with no visible scoring play), not a ledger bug.
PRIME Football does not label every number opponent-adjusted. Public team profiles include season-to-date raw football results such as success rate, pass/rush success, yards per play, explosiveness and tendencies. Net Rating and fields explicitly labeled as adjusted are model-based opponent-adjusted snapshots.
Metrics built from additive weekly counts can be recalculated for a selected range. Model ratings are snapshots and cannot honestly be added or averaged across arbitrary weeks, so PRIME Football keeps those two classes separate instead of pretending every column is rangeable.
PRIME Football uses chronological site weeks. Early opener games that occur several days before the main Week 1 slate are separated into Week 0 so the interface follows how fans actually experience the season instead of forcing every early game into one label.
Weekly matchup pages use the latest overall rating snapshot strictly before the selected game week. That prevents the matchup page from quietly using the result it is supposed to be previewing.
What this means. Some definitions changed during the 2026 season build. Ratings (Net/Off/Def/ST) and SOS/SOR have been rebuilt across every published season (2014–2026) on the same current definition, so those are directly comparable across eras. A few other metrics below are not yet rebuilt historically; treat comparisons involving those as approximate.
Ratings (Net, Off, Def, ST). Every published season (2014–2026) now uses PRIME’s unit-attributed Expected-Points model: garbage-time-excluded possession value plus opponent-adjusted Success Rate and Explosiveness for Offense/Defense, and a separately-attributed Special Teams expected-points unit. Cross-era Net/Off/Def/ST comparisons are directly comparable under this shared production formula; every season was recomputed with the 50% prior-season opponent baseline described above.
SOS and SOR. Every published season uses the current résumé methodology: SOS expresses how difficult the same opponents and locations would be for an average FBS team, on PRIME’s neutral-site rating scale; SOR is actual wins above that average team’s expected wins against the same schedule.
Explosive rate. 2014–2025 use the original definition. 2026 counts a play as explosive when it is successful and gains at least 15 yards on a pass or 10 on a rush, so levels are not comparable across the boundary.
Predictions estimate the game itself. PRIME’s current production prediction model is market-independent. It uses football performance to estimate the projected winner, expected margin and expected score. Sportsbook spreads and totals are not inputs to that expected-score projection.
Betting estimates value relative to a market number. The betting system uses football information plus the posted sportsbook spread and total, together with PRIME’s independent view of the matchup relative to that market. The ATS model targets whether a side covers the spread; the totals model estimates the game’s scoring result relative to the posted total.
Disagreement is expected. An expected score does not automatically translate into a betting selection. For example, the prediction model can estimate a total near or above the sportsbook number while the separate market-relative totals model still identifies the Under. Those outputs answer different statistical questions, so PRIME does not force them to agree.
Walk-forward testing. Historical test seasons are evaluated using only seasons and games available before the test sample.
Immutable live predictions. Prospective prediction snapshots are timestamped and are not silently regenerated after the game.
No fake probability precision. PRIME Football will not publish a model win probability as calibrated until calibration has actually been validated.
Misses stay visible. The long-term goal is a public prediction archive so users can inspect both correct calls and misses.
Read the research. See why Week 4 matters for opponent-adjusted rankings and our play-level study of what turnovers actually cost in PPA per play.
Check the national ratings, then open the weekly slate to see how the teams in this weekend’s games compare before kickoff.