Why You Can Predict the Olympics but Not the World Cup

Sport · Animated Explainer · 52 seconds

Two scoreboards, one lesson in statistics: a simple GDP-and-population model nails the Olympic medal table but cannot pick a single World Cup champion — because summing a thousand events cancels luck, and one knockout does not.

▶ Premiering soon on Instagram & YouTube

The story

Forecasters can predict the Olympic medal table with uncanny accuracy from little more than a country's GDP and population — yet no formula reliably picks the World Cup winner. This video shows why, side by side.

The Olympics awards around a thousand medals across hundreds of events. A country's total is a sum of many near-independent contests, so luck averages out and raw resources shine through — a GDP-and-population model lands within a few medals of reality (R²≈0.9). The World Cup produces one champion from a single seven-game knockout: there is nothing to average, and a single bad 90 minutes ends everything. We tested a macro model on every World Cup from 1930 to 2022 — scored against the full field of entrants, it picked the actual champion zero times. Same ingredients, completely different predictability.

Note: World Cup model (GDP, population, football experience, home advantage) fitted on all 22 tournaments 1930–2022 and scored against the full entrant field. The Olympic side is illustrated with Paris 2024 medal totals versus a representative GDP+population+host forecast of the kind that reaches R²≈0.9.