How well did the model do?
48 teams. 16 cities. 104 matches played — and Spain lifted the trophy, the model's most likely champion since before the opening match. Every pre-match prediction is now lined up against every actual result in the calibration report.
★ Tournament so far
Full scorecard →
KOR(model 78%) lost → 1-0How the model actually did
Brier 0.132 against a 0.25 coin-flip baseline. Favorites won 65.4%. Zero for three in penalty shootouts. Every prediction logged before kickoff, never edited — and every miss owned.
Explore the archive
Edge Index
→Leaderboard ranking all 48 teams by a composite of travel km, time-zone shifts, venue moves, altitude swings, heat exposure, and rest gaps. Sort by any column.
Strength of Schedule
→Opponent difficulty: avg Elo of group opponents, the Group of Death, and a combined travel-plus-opponents leaderboard.
Predictions
→The model's final state: 10,000-simulation Monte Carlo with all 104 real results locked in, exactly as it stood at the final whistle.
Bracket
→Final knockout tree, R32 through the Final, with every result locked in.
Schedule
→Every group-stage match recorded, grouped by date. The knockout results live on /bracket.
Venues
→Interactive map of the 16 host stadiums. Altitude, expected heat, roof type. Mexico City sits at 2,240 m.
Compare two teams
→Side-by-side: travel logistics, opponent strength, Monte Carlo odds. Pick any two of the 48 teams. Default: the host derby — Mexico vs USA.
Download the data
→The complete record as raw JSON: 104 pre-match predictions, 39 daily model snapshots, every score. Re-run the math yourself — that's the point.
How the difficulty scores work
The Edge Index ranks all 48 teams on travel difficulty using 9 factors: total kilometres flown, time-zone shifts, number of venue changes, altitude swings, expected heat, climate variance, rest days between matches, circadian penalty (kickoff time vs. body clock), and host advantage. Each factor is normalised across the field; higher score = harder schedule (0–100 scale).
Opponent Difficulty is the average Elo rating of each team's three group opponents, also normalised 0–100. The Combined Difficulty averages the two — harsh travel plus tough opponents equals the toughest road.
Advancement probabilities came from a Monte Carlo simulation of 10,000 tournaments: every match sampled from a Poisson goal model (mean driven by Elo difference), with real results replacing simulated ones as the 104 games were played.