WW3 Odds Tracker

Aggregated prediction-market and forecasting probabilities. Definition first. No doomsday theater.

Data snapshot: late August 2026 · Sources checked against public markets

22%
Community forecast: World War Three before 2050
Source: Metaculus (3.3k+ forecasters)

1. Definition Used Here

World War Three resolves Yes if, before 1 Jan 2050, there is a military conflict that simultaneously satisfies both:
  • Involves countries representing in total at least 30% of world GDP (PPP) or 50% of world population in any year the conflict is ongoing; and
  • At least 10 million people (civilian + military) are killed in the conflict.
This matches the main Metaculus resolution criteria. Proxies, limited regional wars, and cyber-only campaigns do not qualify.

Without a crisp definition every news cycle looks like “escalation.” Markets that skip this step are mostly measuring anxiety.

2. Core Probability

Metaculus — Will there be a “World War Three” before 2050?

Community prediction 22%

Long-running question with thousands of forecasters. Team internal median has been lower (~17%). Historical great-power war base rates and nuclear deterrence both pull in opposite directions.

3. Key Flashpoint Markets (Polymarket, late Aug 2026)

These are the most liquid related contracts. None of them is “WW3,” but they are the realistic pathways.

China invades Taiwan by end of 2026 ~4%
China invades Taiwan by Dec 31, 2027 ~12%
U.S. invades Iran before 2027 ~16–17%
Russia invades a NATO country by end 2026 ~7%
Iranian regime falls before 2027 ~8%

Volumes on Taiwan and Iran markets have been substantial (tens of millions). Short-horizon invasion odds remain low.

4. Other Signals

Manifold & smaller markets

Various “WW3 by year X” questions trade in the low-to-mid single digits for the next few years, rising toward the longer Metaculus horizon.

Expert / base-rate views

80,000 Hours and similar analyses put great-power war over multi-decade horizons in the roughly 20–35% range depending on exact criteria — broadly consistent with Metaculus.

5. Methodology (how this page is built)

  • Primary number: Metaculus community prediction on the clearly defined 2050 question.
  • Supporting data: Highest-volume Polymarket contracts on the most plausible ignition points (Taiwan, NATO-Russia, Iran).
  • No composite “risk meter” invented here. Weighting arbitrary indicators creates false precision.
  • Update cadence: Manual snapshot. Live scraping of prediction markets is possible but adds fragility and ToS risk.
  • What is excluded: Crowdsourced “how scared are people” maps, pure media-sentiment indexes, and any model that treats every skirmish as linear progress toward global war.

6. Reality Checks

  • These are market prices and community forecasts, not ground truth. Thin markets and headline-chasing can distort short-term numbers.
  • Nuclear weapons, mutual economic dependence, and the memory of the 20th century still exist. That is why long-horizon probabilities are not 50%+.
  • A site that refreshes every hour and screams “odds rising” is mostly selling anxiety. This one prioritizes definition and transparent sources.
  • If you want live numbers, go directly to Metaculus and Polymarket World Events.