CreditVintage / explicit coverage policy / synthetic

Does this batch meet the rules you specified?

Decision-time feature coverage known as of 2024-10-01, checked against a retained policy. Conditions use exact available / expected counts. Report rates rounded for display never decide pass or fail.

Overall policy result: fail
Rules passed0
Rules failed3
Insufficient data0

Inspect each rule

Thresholds come from the retained policy; choose them for your task. The published example uses illustrative limits. Every configured rule remains visible. Empty or undersized populations are insufficient; they cannot pass by dropping missing cases. Swipe or scroll the table horizontally to inspect all columns.

InputAvailable / expectedResultRequired coverageMinimum applicationsPopulationMonthRuleReason
utilization318/320fail1.020currentallcurrent-utilizationbelow_minimum_coverage
debt_to_income319/320fail1.020currentallcurrent-debt_to_incomebelow_minimum_coverage
prior_delinquencies319/320fail1.020currentallcurrent-prior_delinquenciesbelow_minimum_coverage

Separate three questions

Policy pass: only the named coverage and sample conditions were met. A partial policy or a tolerance below 100% can pass with incomplete data. Existing scoring and distribution commands still require every decision-time input.

Verified evidence: events, policy, counts and files rebuild consistently. A verified bundle can contain a failed policy. Its verifier returns the policy result separately.

Trusted policy and sources: keep the expected policy digest outside this bundle and supply it when verifying. That pins the policy; it does not authenticate data sources or certify credit-risk quality.

Canonical policy SHA-256: 5c34c4e386620d6aa17acec056fd87eeca511fa9cdeec4dc59a43564b95c3bed

No outcomes, model fitting, automatic notifications or changes to scoring rules. Any failed rule makes the overall result fail; otherwise any insufficient rule makes it insufficient. The check command returns 0 only for pass, 3 for failed/insufficient rules, and 2 for invalid inputs or inconsistent evidence.