How we operate, in numbers
We publish quarterly statistics so anyone using our data — compliance teams, plaintiffs' lawyers, journalists, researchers — can evaluate our editorial process the same way they'd evaluate a news desk. These numbers are auto-generated from our internal audit log. Nothing is hand-edited.
All-time
2026-Q4 · this quarter
Trailing four quarters
| Quarter | Added | Re-verified | Status changes | Corrections | Retractions | Source uptime |
|---|---|---|---|---|---|---|
| 2026-Q4 | 0 | 0 | 0 | 0 | 0 | 99.7% |
| 2026-Q3 | 0 | 2 | 0 | 0 | 0 | 99.7% |
| 2026-Q2 | 15 | 65 | 2 | 0 | 0 | 99.7% |
| 2026-Q1 | 38 | 0 | 10 | 0 | 0 | 99.7% |
Every incident requires at least one primary or secondary source. We HEAD-check every source URL every 6 hours and fall back to an archived copy when an upstream goes dark.
Methodology →Every status change, severity update, and re-verification writes a row to an append-only audit table — including who made the change and why.
Editorial standards →When we get something wrong, the correction appears on the incident page, dated, alongside the original text. We never silently edit.
Corrections policy →We honor narrow legal-removal requests (defamation, court-ordered redaction, GDPR Article 17). Quarterly totals are published in this report.
Removal-request form →Numbers update hourly via our editorial audit cron. See system status for ingestion cadence and live source health.