FEA Ratings.

Manual team ratings informed by public evidence, election history, and internal model diagnostics. The automated model is maintained separately in Model Lab.