The edge board.

How these picks are made

Every pick is derived from our Poisson scoring model — the same one that powers the main predictions page. We compute per-match expected goals (λ_home and λ_away) from ELO ratings, team archetypes, and pre-match form, then solve probabilities analytically.

Card and corner probabilities use per-match team averages from WC qualifying, adjusted for match intensity (closely contested fixtures and knockout matches see more cards). When API_FOOTBALL_KEY is configured, those baselines refresh live from API-Football.

Each card shows the best model pick at the top with a confidence bar, plus 2–3 additional plays ranked by confidence.