Shutdown early warning, logged before the fact.
The study schedules a 7-day forecast for 152 national units at 06:00 UTC each day. This page lists published hashes; any alerts will be judged against independently verified shutdown onsets.
Log through . No performance score yet.
Daily commitments
Each committed row records a forecast-file hash and its comparison arms. An arm listed here may still lack a populated forecast; this index does not expose its scores or alert counts.
Scroll the table to inspect every field.
| Date (UTC) | SHA-256 | Arms | Status |
|---|---|---|---|
d48a0cbe739eebb565278b095270e583d0d4763cc7b3cb933296c77c68169b69 |
| committed |
The hash lets a released file be checked against this commitment. It does not independently prove its issue time. Dates identify forecast issue days, not independently verified publication timestamps. Scores stay private until the scoring window closes.
Official planned restrictions
The notice register checks official pages for explicit, dated restrictions. A verified entry is labeled “official planned restriction”; an announcement is not evidence that the restriction took effect.
Register status reported as of
No qualifying notice yet.
A notice appears here only after a reviewer verifies an issuer's explicit, dated restriction.
Some configured pages could not be retrieved. This is not confirmation that no restrictions were announced.
6 configured official sources
- Iraq Ministry of Communications
moc.gov.iq
- Syrian Telecom
syriantelecom.com.sy
- Sindh Home Department
home.sindh.gov.pk
- MTN Sudan
www.mtn.sd
- Sudani
sudani.sd
- Election Commission of India
www.eci.gov.in
How it is scored
Study rules as of .
- Eligible warning: 24 hours to 7 days before a confirmed onset. Same-day signals do not qualify.
- Daily alert budget: the top 5 alerts globally. Misses and false alerts count.
- Confirm an onset with 2 independent sources agreeing within 6 hours, plus independent evidence of deliberate action. Connectivity loss alone is insufficient.
- Keep a dispute log with confirmed, disputed and rejected labels and reasons. Reviewers assess onsets without seeing the alerts.
- At 90 days: an operations read, with no accuracy claim.
- Decision review only after both 180 days and 120 confirmed onsets. The later threshold governs.
Outage alerts are candidates for review. They are not confirmed shutdowns. The commitment log makes the study inspectable; it is not a validated warning service.