Transparent by design
Data and ratings
The complete research archive contains 116 club-seasons and 2,530 player-season records. The live game uses a separately reviewed v5 release: exactly 40 playable packs (33 modern and 7 classic) and 797 players whose identity, team-season and exact role evidence passed the publishing gate.
Playable archive versus research archive
A historical roster may be useful for research while still being unsuitable for the game. Sources that only identify a player as DF, MF or FW stay in the research archive until more exact evidence is found. They are not silently converted into CB, CM or ST, and they are not described as playable.
Primary and secondary roles
Recent roles are derived from official Lega Serie A tactical coordinates, with StatsBomb used as an open-data cross-check for 2015/16. Pre-coordinate classic lineups use a separately stored human-reviewed exact-role map bound to the same player, club and season. No DF, MF or FW label is automatically converted into CB, CM or ST. A primary role needs at least two same-season observations; a secondary role needs at least two and at least 20% of the primary role's observations. Every released role keeps a match reference, source type, confidence and evidence hash.
Independent OVR formula
35% club-season strength + 50% player-season contribution + 15% historical achievement. Club strength is calculated from 45% points per game, 35% goal difference per game and 20% final rank. Player contribution uses available appearances, starts, minutes and goals; unavailable assists or individual awards are recorded as missing rather than invented. The six card attributes are transparent role-based gameplay values derived from OVR, not official player statistics.
Static v5 delivery
The browser first downloads a small manifest. Each club-season is a separate content-hashed JSON file and is fetched only when that pack is drawn; the next deterministic pack may be prefetched. The site never calls a football-data provider while a player is visiting.
Editorial corrections
A documented correction may adjust OVR by no more than three points. The reason and source must be retained. The visual Legendary, Elite, Prime and Heritage tiers are derived from OVR and never change the underlying score.
Simulation integrity
A seeded probability model simulates a complete 20-team league. Fixtures, match events and results are reproducible, while forecasts use an independent random stream so viewing a prediction cannot affect the official outcome. Cup opponents belong to a fixed bracket and eliminated clubs cannot return.