All-round powerhouses
USA, Great Britain, Netherlands, Italy, Jamaica, Canada. Strong nearly everywhere rather than concentrated anywhere.
Start · the idea
You need one number that works for every event
The obstacle to comparing nations across events is that the events do not share units. Seconds, metres and points are not interchangeable, and a country strong in the hammer looks like nothing at all next to a country strong in the 100 m unless you can put both on one scale.
World Athletics already publishes that scale. Every mark in the top lists carries a WA point score designed to be cross-comparable. Whether the tables are perfectly fair is its own question — one I took apart separately in Auditing the Scoring Tables — but as a common currency they are the best available, and they are what the sport itself uses.
↓ the map
Five archetypes, and one enormous outlier
Clustering the 43 nations on their strength across sprints, middle distance, distance, jumps and throws produces five groups that need almost no interpretation:
Distance specialists
Kenya, Uganda, Ireland. Elite at distance and middle distance, and pointedly narrow outside it.
Throw & jump powers
Poland, China, Cuba, Nigeria and others whose strength sits almost entirely in the field events.
Distance & middle powers
Australia, France, Germany, Norway, Japan, Spain. Broad established programmes with an endurance lean.
Emerging / single-event
Strong in one niche — a Bahamas 400 m, a Trinidad sprint — without the breadth behind it yet.
And then the USA
On PC1, overall depth, the USA sits far beyond every other nation. It is not the best of a group, it is its own region of the chart.
↓ the underlying matrix
Where each country's points actually come from
Behind the clusters sits the thing they were computed from: a heatmap of every nation against every discipline group. The specialists are obvious as bright horizontal bands, and the all-rounders as nations with no dark cells at all.
↓ what this cannot tell you
Absolute strength, not efficiency
This measures how much elite athletic capital a country has, not how impressive that is given its size or wealth. On an absolute matrix, large populous countries dominate almost by construction. The genuinely interesting per-capita and per-GDP overperformers — Jamaica, Norway — are a natural next layer, and they are not in this version.
Two modelling choices are worth stating plainly. Strength per nation and discipline is the mean of that nation's top-5 WA scores in the group, which is deliberately robust to a single freak performance but also means a country with exactly one world-beater scores modestly. And a nation absent from a discipline entirely is filled to the column minimum before z-scoring, treating absence as no elite presence. That is a defensible reading, but it is an imputation and it shapes the clusters.
A re-fetch of the raw top lists once produced different clustering values than the committed table, and the cause was never isolated. The published numbers are the ones in this repository; that discrepancy is recorded rather than quietly resolved.
Finish · how it was built
Where the numbers come from
- Data
- World Athletics top lists, scraped with mark, WA points and nationality
- Sample
- 6,400 performances, 16 events, 2023–24, 43 nations
- Strength metric
- Mean of a nation's top-5 WA scores per discipline group, then z-scored
- Clustering
- k-means over the five discipline groups; PCA for the map
- Output
national_profiles.csv— every nation's per-discipline strength and archetype- Built with
- Python, pandas, scikit-learn, matplotlib
- Companion
- Auditing the Scoring Tables, which tests the currency this analysis spends