#world-cup
2026
- How the five AIs actually did
The Group Stage of the 2026 World Cup finished today, so it’s time to mark the homework. Back in the launch post five AI models, a deterministic lookup table, and one football fan predicted the same 72 fixtures before a ball was kicked. I’ve been scoring the results on AIWC26 each day as they came in. I’ve got no real opinion on the teams; I’m interested in what the scoring reveals about the models.
- Inside the BBC's 3D World Cup viewer
While updating scores for AIWC26 this week I came across BBC Sport’s 3D World Cup viewer, which launched on 12 June for matches broadcast on the BBC. Every match played so far is sitting there to replay in full, not just the live ones. It’s UK-only, it’s a beta, and somehow this is my second football post in a row, despite still having zero interest in the sport itself. The engineering underneath it is another matter.
- Five AIs predict the World Cup
The World Cup rolls around every four years, and with it the office sweepstake, the ritual where everyone gets handed a team at random and pretends to care how it does. I have no interest in football (I had to read a beginner’s guide to how it all works just to follow what’s going on), and I definitely don’t know enough to make an informed prediction of my own. So I thought I’d hand the job to five AI models instead, get them to commit their predictions in public, and keep score. The result is AIWC26, a page that shows every group-stage prediction from five models side by side, scores them against the real results as they come in, and ranks the models in a league table. It looks like Ceefax, because of course it does.