Research · Counter-narrative
Where the model and FIFA rank disagree
Snapshot · 2026-05-22Model 1.0.0For each WC 2026 nation we compare the model's tournament-winner probability against a FIFA-rank-implied baseline. The implied probability is a transparent monotone transform of FIFA rank, a softmax over the negative rank with temperature τ = 8 so a higher rank gives a higher implied probability and the 48 implied probabilities sum to one across the field. The table below lists the teams with the largest positive and negative deltas. Methodology in /docs/methodology/.
48 teams · transform: implied_j = exp(-rank_j / tau) / sum_k exp(-rank_k / tau); teams with no FIFA rank substitute the field's median rank
Top 10: model rates above FIFA-rank baseline
Largest positive delta: Argentina (+9.06pp)
Per-team delta between the model's tournament-winner probability and a FIFA-rank-implied baseline. The Standard Pass unlocks the full table.
24h money-back, no questions asked
Top 10: model rates below FIFA-rank baseline
Largest negative delta: Belgium (-2.95pp)
Per-team delta where the model rates the team below the FIFA-rank-implied baseline. The Standard Pass unlocks the full table.
24h money-back, no questions asked
Standard Pass
Unlock the counter-narrative scoreboard
The Standard Pass unlocks the full table plus every other gated section across the site, valid through the tournament.
Every forecast graded against the real result, scored on 987 matches since 2014. See the scorecard.
24h money-back, no questions asked·No subscription, no auto-renewal·Access through 31 Dec 2026. See refund policy.
What this comparison is, and isn't
The implied probabilities here are derived solely from each team's public FIFA rank via a documented softmax transform. This page does NOT compare to bookmaker odds, to prediction-market prices, or to any external commercial probability source. The only baseline used is FIFA rank.
A large positive delta is not a recommendation; it is a description of where the model's ensemble (Elo + Dixon Coles + Hierarchical Poisson, calibrated by isotonic regression) reaches a different conclusion from what FIFA rank alone would suggest. The model's out-of-sample calibration is published on the methodology page; FIFA rank's implied probability is shown here for comparison only, not validated as a forecast.