How fresh is the data behind a censorship verdict? A per-country quality audit

An observatory that asks you to trust its censorship verdicts owes an honest answer to a prior question: how good is the data behind any one verdict? It is not uniform. Voidly has evidence for 215 countries but recency varies by five orders of magnitude — 13 measured within 24h, 174 within a week, but 28 NOT measured in over 30 days (Djibouti's latest is 158 days old, on 2 measurements; a "high-risk" verdict there is a statement about the past). Recency is not the only axis: over 30 days, 54 countries are "bursty" (signal arrives in spikes, not a steady stream) vs 24 stable. Combining freshness x volume x stability x source-diversity into a composite 0-100 data-confidence score: 53 countries high (>=70), 52 medium, 82 low (<40). Iran scores 74 (fresh, ~3,500 measurements / 3 sources — bursty days outweighed by volume + corroboration); most low-confidence countries are small/rarely-probed (real signals, wider error bars). A LOW score does NOT mean uncensored — it means weight with caution. Honesty cuts the other way too: of 5,677 measured domains only 0.5% carry a Citizen Lab content category, so any "by category" claim is a floor not a census. All of this is now live + machine-readable (GET /v1/measurement/freshness, /coverage, /quality, /category-coverage) so the quality of the data travels with every query. The same standard Voidly applies to its forecasts and classifiers — report the honest number, name the limitation — now applies to the measurement layer underneath them.