Deep dive — chart 1

Chart 1 — rolling mean underwriting-risk-score by applicant group

The chart your Chief Compliance Officer should look at first. Where the silent H3 divergence begins, in one visual.

7-day rolling mean underwriting-risk-score per applicant group over 90 days

7-day rolling mean underwriting-risk-score per applicant group over the 90-day audit period. H3 divergence begins Day 44 and never recovers.

What you are looking at

X-axis: Day 1 through Day 90 of the audit period
Y-axis: 7-day rolling mean of the AI's underwriting-risk-score (0-100 scale)
Each line: one applicant group (H1 = prime + majority-white zip, H2 = prime + mixed zip, H3 = 30-50 near-prime majority-minority zip, H4 = first-time homebuyer, H5 = self-employed)
Baseline window: the first 30 days establishes what "normal" looks like per applicant group

What the eye sees

Four applicant groups stay roughly parallel across the 90 days. One applicant group — H3 — rises. The rise begins around Day 44, and it does not recover. The line separates from the pack, moving into the decline / non-QM zone.

This is the pattern an experienced fair-lending officer instinctively recognizes: something in the H3 stream changed, and the AI is scoring it as higher risk than baseline. The 47-percentage-point lane-shift consequence is the follow-on.

What the sensor fires on

The sensor is not looking at the eye's pattern. It is running four quantitative thresholds against each 7-day window per applicant group:

Mean-shift: is the 7-day mean > 1.5 standard deviations from baseline mean?
Variance-ratio: is the 7-day variance > 1.5x baseline variance?
KL-divergence: is the shape divergence between baseline and current distribution > 0.2?
Group-differential: is the max cross-group shift > 12%?

Day 44 fires H3 KL-divergence 0.833 (4x the threshold). By Day 65 it hits 0.963. The pattern the eye sees is the pattern the sensor quantifies.

What regulators care about

Regulators do not read charts — they read attestations. But when an attestation is challenged, the chart is the exhibit. CFPB supervisory examiners, DOJ fair-lending referrals, HUD complaint investigators, state banking-department examiners, and plaintiff's expert witnesses all understand the "one applicant group drifts silently" pattern instantly.

The chart also tells the counterfactual story: had this monitoring been in place operationally, the drift window would have been ~7-14 days rather than 45+.

The next chart

The rolling-mean chart shows when. The baseline-vs-drifted chart shows how the H3 distribution shape shifted.

Chart 2 — baseline vs recent distribution →

$499 Snapshot. 3 business days.

Same 3-chart set for your lender's actual AI surface + full analysis report.

Buy $499

Back to housing sample hub.