How the model works
Stats drive the model
Every player-season is normalized within its era and position, so a 1990s guard and a 2020s guard are judged on the same scale. Production and efficiency carry the most weight; awards break close ties and raise the ceiling when the production backs it up.
Roster construction matters
Famous names are not enough. The model rewards a real primary creator, floor spacing, rim protection, ball security, and rebounding — and penalizes redundant ball-dominant scorers, clogged spacing, and missing interior defense.
The 40-game gauntlet
40-0 is a standardized perfect-season path: 29 regular-season games, a 4-game conference tournament, and a 7-game national tournament that escalates to elite opponents. Even a monster team can trip late — winning three straight elite opponents to finish is genuinely hard.
A projection, not a claim
A 40-0 result is a calibrated model projection, not a historical claim. It is built to start a fair debate, not end one. Every result stores its model and data version, so older shared results stay valid under the version they were scored with.