update-pairwise-rating

Purpose

Update latent candidate ratings and uncertainty from a pairwise outcome using an explicitly chosen rating model.

Input contract

required: [prior_ratings, pairwise_outcome, rating_model, update_parameters]
optional: [rating_history, draw_policy, uncertainty_floor]
constraints: [model, outcome direction, confidence, and parameter values are recorded]

Procedure

1. Validate the pairwise outcome and retrieve the two prior rating states.
2. Apply the declared Bradley-Terry, Elo, Glicko, TrueSkill, or equivalent update equations.
3. Update both ratings and uncertainty, preserving the pre-update state and the reason for the magnitude.
4. Emit the updated ratings, uncertainty change, and consistency flags for downstream audit.

Output contract

produces: [updated_ratings, uncertainty_updates, update_trace, consistency_flags]
delta_fields: [findings, evidence_updates, uncertainties, decisions]