How the PR Score is built.
Let's score Felix Ojo.
Below, we walk through each of the five components in order. For every one, we show how it was calculated, what the inputs were, and how it compares to the rest of the class.
Weighted Service Average
Each of the four services (ON3, 247, Rivals, ESPN) publishes a national rank. We translate each into a 0–50 score using their published rank-to-score curve, then weight by historical predictive accuracy.
Rank Momentum
We compute the slope of each service's rank for this recruit over the trailing 90 days. Positive slopes earn points proportional to magnitude and consistency across services. Recruits with no movement earn the midpoint.
Position Scarcity
NFL Draft position values × HS supply. If the consensus top 100 is structurally short on a position relative to NFL demand, blue-chip recruits at that position carry a premium. Re-fit annually using the prior 10 drafts.
Cross-Service Stability
Standard deviation of the recruit's rank across the four services. A tight cluster scores near 10; wide divergence scores near 0. Acts as a confidence multiplier on the rest of the model — disagreement should reduce certainty.
Elite Recognition
Binary or graded credit for invites to combine events, Polynesian Bowl, All-American Bowl, Five-Star Plus+ designation, and individual service "5★" status. Capped at 10 to prevent the most-decorated recruits from running away with the score.
The composite is a snapshot, not a prophecy.
The PR Score answers one question: based on everything observable about a recruit today, how does this player compare to others in the same class? It is not an NFL projection, a draft model, or a guarantee of college performance.
We publish error bars where they exist. We disclose every adjustment. If the model is wrong about a player, we want to be specific about how.