Rate Authority.

PolicyChat Methodology — Index

Updated 2026-05-21

PolicyChat Methodology — Index

Quick links to every PolicyChat methodology page. Each page reads as a research Methods section — data sources, analytic procedure, fidelity declarations, and known limits.

Underlying engine

PolicyChat is built on the chorus_stage2 methodology — a Bayesian conviction- filter framework originally developed for sports / commodity / political prediction modeling. Cross-application to insurance applies the same |p−0.5|

0.20 publish gate to magnitude claims.

Errata

Methodology pages are versioned. Material changes are dated. Errata at [email protected].


Last updated: 2026-05-21.