13 stats that prove the USMNT will win the 2026 World Cup - ESPN

Gerald Skinner used to work for NASA. He wrote papers with titles like «The Galactic distribution of the 511 keV e+/e- annihilation radiation» and «Antimatter in the universe and the PAMELA/FERMI/AMS anomaly.»

But in 2009, he turned his attention to something a little more frivolous and presumably much easier to understand than the behavior of the invisible forces that shape the Milky Way: the game of soccer.

Along with a colleague, Guy Freeman, Skinner wrote a paper for the Journal of Applied Statistics to «consider what, if anything, can be deduced from the result of a match about the relative strengths of the teams.»

Their answer is one of my favorites in all of academic literature. Skinner and Freeman make the claim that if the result of a soccer match tells us the true strength of each team, then we should never see a scenario in which Team A beats Team B, Team B beats Team C, and Team C beats team A — what they call an «intransitive triplet.» If the scoreline is truly representative of who the better team is, then this would never happen.

They then looked at every set of triplets, where three teams play each other, over the course of World Cup history from 1938 through 2006, and found that 12% followed the «intransitive» pattern they said shouldn't happen. That percentage might seem low — until you realize that if the results of matches were completely random, we'd expect intransitive triplets to occur 25% of the time. The results of the World Cup were only slightly better than random.

So discouraged were the authors by their findings that they titled the paper: «Are soccer matches badly designed experiments?» They go on to say this about previous tournaments: «Even on very optimistic assumptions, there is less