Build the pipes.Then print the edges.

Start with structure, then roles, then lines and projections. No fake confidence while the inputs are still forming.

Real schedule data only. Player projections stay off until role, usage, injury, and market inputs pass validation.

These are the first markets worth supporting because they connect cleanly to usage, role, game environment, and sportsbook lines.

The first JSON files are placeholders by design. They tell every future page and script where the NFL truth should live.

Football is role-driven. The first version should handle unknowns cleanly instead of pretending camp noise is a finished model.