World Cup 2026 Predictions
ELO-based match probabilities w/ 10,000-simulation Monte Carlo tournament model for all 48 nations.
Tournament Winner Probability
Percentage of 10,000 simulations each nation wins the tournament outright.
Why do these numbers change on refresh?
Refresh this page and the probabilities will shift. Argentina might go from 19.4% to 20.1%. England might drop half a point. This is intentional.
The model runs 10,000 simulations of the entire tournament on every page load. Each simulation samples from the probability distributions we've built — it doesn't pick a single "most likely" outcome and report it back to you. It plays the whole tournament ten thousand times and counts how often each team wins.
Nassim Taleb's central argument in Fooled by Randomness is that we systematically confuse the map for the territory — treating a model's output as a fact about the world rather than one draw from a distribution of possible worlds. A number that subtly shifts on each visit is a reminder that 19.7% is not Argentina's probability of winning. It's our best estimate of it, with all the uncertainty that implies.
A number that never changes signals false confidence. A number that moves slightly on each visit is the model being honest about what it knows and doesn't know. The instability is the information.
The variance between refreshes is also a rough proxy for uncertainty. Groups that are genuinely competitive produce wider swings — a qualify probability bouncing between 58% and 64% is telling you something the headline figure doesn't: this group is close, and the outcome is sensitive to how a few 50/50 matches fall. We could run 100,000 simulations and pin the numbers down tighter. We've chosen not to. The flicker is an honest signal of a genuinely uncertain world.
Group Stage Overview
Qualification probability per team across all 12 groups. Click a group for fixtures and match predictions.