WAR DETECTOR

Methodology · version 2026-11-v2

How the War Detector Index is calculated

The index runs from 0 to 100 and sums up how close the international situation is to escalating into a global conflict. It is a statistical indicator built from public data, not a real probability or a forecast. It is recalculated every hour.

1. Risk of each conflict (0 to 100)

Each active conflict gets a score, the weighted sum of three factors:

2. Global index

Each component below adds points. The total goes through a saturation curve, 100 × (1 − e−points/60), which keeps the result between 0 and 100. Every country that goes to war always raises the index, but the number never exceeds 100.

ComponentRuleNow
Countries at war+1 per country44
Nuclear powers at war+4 per power20
Permanent members of the UN Security Council+3 per member6
Defense alliance members at war+5 per country10
Nuclear powers on opposite sides+120
US, Russia or China on opposite sides+250
Continents affected+1.5 per continent9
Average conflict riskup to +102.72
Total points → index91.72 → 78.3

An alliance counts as triggered when a member fights a country outside it directly and other members are still out of the war and could be dragged into it. Each member country at war adds points once, even if it belongs to several treaties: the US at war counts as a single source of escalation. The lists of nuclear powers, permanent members and alliances are reviewed by the team.

3. Bands

4. Sources

5. Limitations

Conflict data arrives late and may be revised. GDELT measures what is reported, not what actually happens, so regions with little press coverage show up less. The weights are editorial choices, published here and versioned. Any change creates a new version.