Data sources & methodology

This page lists where each figure shown on WorldRelat comes from, how often it's refreshed and what its limits are. It documents the source and update cadence by type of indicator, not the current value of every single figure -- those live on each country and relation page itself. See the About WorldRelat page for the story behind the project.

Summary table

Population

Source: World Bank Open Data. Coverage: 2025 for most countries (a few territories have no World Bank figure; an older figure of unknown year is kept). The reference year is stored with every figure and shown on the War Simulator result. Totals follow the World Bank series: Moldova excludes Transnistria, and China excludes Hong Kong, Macau and Taiwan.

GDP

Source: World Bank Open Data. Coverage: latest available (2025 for most countries, older for a few, back to 2011).

Military expenditure

Source: World Bank Open Data. Coverage: latest available (2024 for most countries; missing or older for some).

Military personnel (active + reserve)

Source: Wikipedia (aggregating IISS Military Balance figures). Coverage: third-party data, not official, curated by WorldRelat and updated every year as new figures are published. Unlike the World Bank figures above, this isn't an official government statistic -- it's a third-party aggregation we use as the best widely-available estimate, and it was collected once rather than refreshed on a schedule.

Influence score

Source: WorldRelat methodology. Coverage: updated automatically on every scraper run. A country's influence score measures how much documented international activity it has. For every relationship the site tracks, the length of the Wikipedia article about that pair is recorded, and a country's influence is the sum of those lengths across all of its relationships. That total is then shown on a 0-10 scale relative to the most influential country in the dataset. In other words, it reflects how extensively a country's foreign relations are documented -- which tends to go together with, but is not the same as, economic or military power. It does not use population, GDP or military figures.

Relationship score

Source: WorldRelat relationship methodology. Coverage: updated automatically on every scraper run. Every relationship score starts from the Wikipedia article dedicated to that pair of countries. The article's introduction and its full body are read separately by VADER, an open-source sentiment analysis tool, and combined into a single score out of 10, weighted toward the introduction. When the article is short, the score is pulled toward a neutral middle so a stub can't produce an extreme verdict. Shared membership in the European Union or NATO nudges the score up slightly. The final number is sorted into five bands, from Excellent to Very bad. Visitor votes never change this score. See About WorldRelat for the full explanation.

How the War Simulator uses these indicators

The War Simulator combines five indicators for every country on each team: influence score (30%), military expenditure (25%), GDP (20%), military personnel (15%) and population (10%). Each factor is shared between the two sides according to the square root of its value, so a very lopsided matchup doesn't collapse to exactly 100% against 0%. If an indicator is missing for both teams it is left out of the calculation; if it is missing for only one team, that team counts as zero on that factor. Alliances, nuclear weapons and geography do not change the percentage. The result is a structural comparison of countries, not a prediction of a real conflict.

Reach, staying power and strategic capabilities

Below the win probability, each War Simulator result shows three context sections. They help read the matchup, but none of them changes the percentage, which still uses only the five indicators described above. The military power ranking uses exactly the same formula, with each country matched one on one against every other.

Reach & logistics. The distance is the great-circle distance between the capitals of the two closest countries on opposite sides, calculated from capital coordinates; land borders come from the neighbour list of each country. The power-projection class (territorial defence, regional reach, expeditionary or global) is a hand-curated grouping based on public reporting. For countries without a curated entry, the class is inferred from defence spending: regional reach from US$20 billion a year, territorial defence below that.

Staying power. The index (0–100) averages up to six indicators of how well a country could endure a long war: energy imports as a share of energy use, food imports as a share of merchandise imports, foreign reserves in months of imports and the latest annual GDP growth (all World Bank), the arms trade balance (imports against imports plus exports, SIPRI trend-indicator values distributed by the World Bank) and defence spending as a share of GDP. Each indicator is scored by ranking the country against all countries with data (100 = better than all of them, 50 = world median). For a team of several countries the value is the GDP-weighted average. These World Bank figures were downloaded in October 2026, use the latest year available for each country and are not refreshed automatically, so the year can differ between countries.

Strategic capabilities. Six categories (aircraft carriers, nuclear-powered submarines, fifth-generation fighters, intercontinental missiles, military satellites and armed drones) are shown as tiers from 0 to 3. The tiers are curated by hand from public sources, mainly the IISS Military Balance, SIPRI, national defence ministries, the Union of Concerned Scientists satellite database and specialist press. They are approximate groupings, not exact inventories; they describe what exists, not how well it would perform, and a country with no entry counts as tier 0.

Update frequency

Population, GDP and military expenditure are refreshed from the World Bank API (last full refresh: 5 October 2026), and the year of each figure is stored with it and shown for every country. Military expenditure reaches the World Bank from SIPRI about a year late, so most countries show 2024 data until the next release. Military personnel figures are reviewed and updated once a year, as new figures are published. Relationship and influence data are recalculated each time the scraper runs, and for every pair the most recent run is the one displayed. There is no fixed publication calendar.

Data limitations

• Wikipedia coverage is uneven. Countries and relationships that have more written about them score higher on influence, so small or less-documented countries are under-represented regardless of their real importance.
• Sentiment analysis reads wording, not events. A calmly written article about a tense relationship can score better than the relationship deserves, and the opposite can also happen.
• Military personnel figures are a third-party aggregation of different years, not an official count.
• Some countries have no GDP or military expenditure value, or only an older one. The site shows only what exists.
• Taiwan, the Cook Islands, the Holy See and Niue are not in the World Bank population series, so they keep an older population figure whose year was not recorded.
• The dataset covers 204 countries; international organisations and other non-country entities are excluded from rankings and pages.
• The War Simulator is hypothetical. The percentage does not account for terrain, morale, intelligence or political goals; reach, staying power and strategic capabilities are shown as context and do not change it.

Corrections and contact

WorldRelat is maintained by Jordi Fuster, a statistician. If you find a figure that looks wrong or want to suggest a better source, please write through the contact page; every report of incorrect data is looked into.