01 / OBSERVED RECORD
What happened?
GDELT encoded Destroy property, Use conventional military force, not specified below. Involved actors: Russian, Ukraine.
Location: Kyiv, Kyyiv, Misto, Ukraine
Approximate GDELT action-location centroid; not a verified affected facility. This story contains 2 coded locations; the map uses one representative location.
These are reporting counts, not independent confirmations, casualties or disruption quantities.
Open original source ↗02 / MULTI-LABEL CLASSIFICATION
One story. Several dimensions.
Event types
Transmission mechanisms
Potentially exposed assets
03 / CONDITIONAL BUSINESS EFFECT
What could it disrupt?
Asset damage → Crude oil & refining → Production / delivery availability risk
Why this link exists: CAMEO 1712 · Asset terms in URL: oil
Conditional pressure estimate; not verified physical loss or market-price impact. Evidence: MODERATE RULE MATCH.
Show model factors
- event type severity
- 100.0
- normalized evidence
- 0.692
- mechanism strength
- 0.92
- target sensitivity
- 0.92
Conditional pathway, not proof of an outage, commodity price move or company loss.
Watch next: Verify facility identity, lost output, inventory buffers and observed price movement.
Armed disruption → Crude oil & refining → Production / delivery availability risk
Why this link exists: CAMEO 190 · Asset terms in URL: oil
Rule: blackdeck-taxonomy-2.0/asset:crude-oil/armed-disruption
Conditional pressure estimate; not verified physical loss or market-price impact. Evidence: MODERATE RULE MATCH.
Show model factors
- event type severity
- 100.0
- normalized evidence
- 0.692
- mechanism strength
- 0.88
- target sensitivity
- 0.92
Conditional pathway, not proof of an outage, commodity price move or company loss.
Watch next: Verify facility identity, lost output, inventory buffers and observed price movement.
04 / WHAT IS ACTUALLY NEW?
Escalating signal
2 similar prior source-story groups in 7 days; similar reporting on 5 dates in 14 days. This is a coverage comparison, not a count of verified recurring incidents.
Retained new-information weight: 89.4%. Reporting persistence: 43.1/100 (coverage index only).
The model restored some attention because the coded severity or matched mechanism/asset changed. This change still needs source verification.
Repeated reporting reduces editorial novelty. It does not reduce verified physical damage or establish that a disruption remains unresolved.
Inspect the formula and limitations →