About
What ScoutIndex is, and where the numbers come from
ScoutIndex™ is an NFL statistics database — every team, player, game and (for recent seasons) every play since 1999 — built on a sport-agnostic core schema so other leagues can be added without a rewrite. It is an independent project and is not affiliated with or endorsed by the NFL.
Data source
All data comes from nflverse, a community-maintained, MIT-licensed set of NFL datasets (play-by-play with EPA/WPA from nflfastR, weekly player stats, rosters, the draft, the combine, snap counts). No unlicensed scraping and no paid feeds are used. Facts such as scores and statistics are not themselves copyrightable; the ScoutIndex compilation, schema and any derived ratings are its own.
Freshness
Scores and win probability refresh every 15 minutes; the full stats/ratings pipeline runs every 3 hours. Live job-by-job status: Data status →
Ratings & win probability
The ScoutIndex Rating (v2) is live — a transparent, position-relative 0–99 score for QB / RB / WR / TE / K plus a defensive impact score, built from nflfastR EPA and counting stats, with every input and weight published. Each rating carries a model_version; a formula change ships as v2 rather than overwriting history. QB/RB/WR/TE ratings also come with a FIFA/Madden-style attribute card breaking the season down into named, formula-backed sub-scores.
ScoutIndex Win Probability (v1.5) is a team Elo model plus a backtested starting-QB overlay — published and tracked on Methodology.
Still to come: opponent adjustment for the Rating; an ensemble win-probability model with a proper rest/travel and weather overlay.