Module effect 

Effect estimation from posterior samples (spec §5.2).

This module computes effect estimates from the 9D posterior over quantile differences. The primary metric is max_k |δ_k| - the maximum absolute effect across all deciles.

Effect Reporting (spec §5.2)

• max_effect_ns: Posterior mean of max_k |δ_k|
• credible_interval_ns: 95% CI for max|δ|
• top_quantiles: Top 2-3 quantiles by exceedance probability

Functions

compute_effect_estimate

Compute effect estimate from delta draws (spec §5.2).

compute_effect_estimate_analytical

Compute effect estimate from posterior mean and covariance (analytical).

compute_top_quantiles

Compute top 2-3 quantiles by exceedance probability.

regularize_covariance

Apply variance floor regularization for numerical stability (spec §3.3.2).