College Football. Decoded.
Rivals Rundown is built to answer a few deceptively simple college-football questions: How good is the team? What does its schedule mean for the season? What does the model expect in each matchup? and Where does that view differ from the market? The site publishes enough methodology to make every number understandable and auditable without publishing the proprietary formulas, weights and calibration logic that create the model.
How the pieces connect
- Team strength: RPR provides the national view of underlying team quality.
- Schedule context: each team is evaluated against the opponents and conditions on its actual schedule.
- Game projections: the matchup view produces projected scores, margins and win probabilities.
- Season outlook: game-level outcomes roll into projected wins and record distributions.
- Market context: preseason market information is compared with the finished football projection as a separate layer.
Important: market information does not change RPR, projected wins, projected scores, margins or win probabilities.
Core rankings and season definitions
Rundown Power Rating (RPR)
RPR is Rivals Rundown’s measure of underlying team strength on a neutral field. It is designed to separate how good a team is from the schedule it happens to play. Higher RPR means a stronger underlying team rating.
RPR Rank
RPR Rank orders every FBS team from strongest to weakest by RPR. This is the primary national power ranking on Rivals Rundown.
Projected Wins
Projected Wins is the model’s expected regular-season win total after team strength is run through the team’s actual modeled schedule. It is an expectation, not a prediction that one exact record must occur.
Projected-Wins Rank
Projected-Wins Rank orders teams by expected wins. It can differ significantly from RPR Rank because schedule difficulty and game context affect the path to a final record.
Schedule Strength Rank
Schedule Strength Rank describes the relative difficulty of the opponents and schedule environment a team faces. It helps explain why two similarly rated teams can have very different projected records.
Record Distribution
The record distribution shows the range of season outcomes produced by the model rather than presenting one expected-win number as certainty. The most likely record and central outcome range provide context around the average.
Game projection definitions
Projected Score
The projected score is the model’s game-level scoring expectation for both teams after team strength and matchup context are considered. It is best read as a center point for the matchup, not an exact-score promise.
Projected Margin
Projected Margin is the expected scoring difference between the two teams. Positive or negative presentation depends on the team and page context, but the magnitude represents the model’s expected separation.
Projected Total
Projected Total is the combined scoring expectation for both teams in the matchup.
Win Probability
Win Probability translates the game projection and its uncertainty into the estimated chance that each team wins. The probabilities are calibrated so that ordinary favorites are not treated like certainties simply because the projected margin is positive.
Score Confidence
Score Confidence summarizes how stable the inputs behind a game projection are. It is a guide to projection certainty, not a separate prediction and not a guarantee that a high-confidence game will finish near the projected score.
Team-information and uncertainty definitions
Roster Fragility
Roster Fragility describes how sensitive a team’s outlook may be to uncertainty around areas such as quarterback, continuity, depth, transfers and coaching transition. Higher fragility means the projection deserves more caution.
Quarterback and roster context
The model considers relevant roster and quarterback context as part of team evaluation. Public pages describe the resulting outlook and uncertainty; Rivals Rundown does not publish the proprietary weights assigned to individual inputs.
Venue, travel and rest
Game projections can reflect where a game is played and the broader scheduling environment. These contextual factors are applied at the matchup level rather than being treated as permanent team strength.
Market definitions
Market Total
Market Total is the posted preseason regular-season win-total benchmark used for comparison with the model.
Market Price
Market Price reflects how the market is pricing one side of a posted total. Price can contain information beyond the headline number itself.
Effective Market
Effective Market is Rivals Rundown’s estimate of the market’s price-aware center. It allows a heavily priced total to be compared more fairly with the model than the posted number alone. The exact price-to-center transformation is proprietary.
Adjusted Edge
Adjusted Edge is the difference between the model’s market-comparable projected wins and the Effective Market. Positive values indicate the model is higher than the market center; negative values indicate the model is lower.
Market Basis
Market Basis identifies how many games are included in the market-comparable season projection. This matters when the modeled full schedule and the schedule covered by a posted win total are not identical.
Signal
Signal summarizes the strength and reliability of a model-versus-market disagreement after uncertainty is considered. Strong represents a larger disagreement with sufficient model support. Lean represents a moderate disagreement. No Edge means the model and market center are relatively close. Review identifies a meaningful apparent disagreement that carries elevated uncertainty and deserves additional scrutiny.
The exact thresholds and weighting used to assign signal tiers are proprietary. Signals are model context, not wagering recommendations.
What goes into the football model
At a high level, Rivals Rundown evaluates information across several football dimensions:
- underlying team power and performance quality
- offense, defense and special-teams strength
- quarterback outlook and experience
- returning production, continuity, depth and roster change
- coaching stability and transition
- transfer-portal additions and losses
- opponent quality and schedule difficulty
- venue, travel, rest and sequencing
- matchup-specific scoring environment
- uncertainty and projection stability
Not every factor is treated equally, and the importance of an input can depend on context. Rivals Rundown does not publish the proprietary weighting system, source-blending rules or internal adjustment formulas.
Rundown Game Lab
Game Lab is a scenario-testing environment built from the same national team-strength framework. It lets users explore hypothetical FBS matchups and change selected conditions such as venue and available scenario inputs. Game Lab applies those choices to the matchup projection while protecting against false precision in extreme scenarios.
Game Lab scenarios are temporary. They do not change official RPR rankings, team pages, schedules, projected wins or weekly projections.
Weekly Projections
Weekly Projections reorganize the official game-level model into the current slate. They are not a separate model. As schedule details and team information evolve, the weekly view can be refreshed while preserving the same definitions used across team pages and National Rankings.
Season-specific schedule treatment
When a conference schedule includes unresolved, flexed or otherwise nonstandard games, Rivals Rundown keeps the full football projection and the market-comparable projection clearly separated. For the current 2026 Pac-12 structure, the full model includes the projected flex matchup while the preseason win-total comparison uses the games covered by the posted market total.
Model and market snapshots
Model version identifies the football-model framework. Site data updated identifies when the current football dataset was generated. Market snapshot identifies the date of the market information used for comparison. Keeping these dates separate prevents a market refresh from being mistaken for a change to the underlying football model.
What we publish—and what remains proprietary
Rivals Rundown publishes definitions, model scope, interpretation guidance, coverage counts, uncertainty labels and the conceptual relationship between team strength, schedules, game projections, season outcomes and market comparisons.
We do not publish proprietary source weights, coefficients, calibration constants, internal thresholds, reconciliation formulas, price transformations, simulation tuning, data-blending rules or other implementation details that would reproduce the model. Transparency should help readers understand the output without turning the methodology page into the recipe.
Corrections and updates
College football changes quickly. Quarterback decisions, injuries, suspensions, transfers, depth-chart changes, coaching developments, schedule adjustments and venue information can alter the inputs available to the model. Rivals Rundown may update projections when meaningful new information is incorporated. A correction to bad data is different from a model-methodology change, and the site aims to keep those concepts separate.
How to interpret Rivals Rundown
RPR, projected wins, record distributions, projected scores, probabilities and Adjusted Edge are estimates—not guarantees. College football contains injuries, turnovers, weather, game-state decisions and other randomness that no pregame model can eliminate. Use the numbers as a structured way to compare teams, schedules and scenarios rather than as certainty.
Market information is provided as analytical context and is not a wagering recommendation. Market totals and prices can change after the displayed snapshot.