Sharpest call · AJ Auxerre
Expected value: +$8.54 on a $20 stake.
Expected value: +$8.54 on a $20 stake.
Joint prob 5% lower than naïve — books overprice this kind of pair.
Expected value: +$3.81 on a $20 stake.
Market: Pinnacle ·
Olympique Lyonnais scores shown first. Bivariate Poisson with team xG from recent form — most-likely score ≠ most-likely winner.
Probabilities derived from a Poisson model on expected goals. Edge % shown only when the book posts both sides of a market.
Joint probability comes from the score grid (the truth). Naïve is what a book gets by multiplying legs — they undershoot correlated positive legs like "home + BTTS" and overshoot anti-correlated ones like "draw + over 2.5". Buy positive correlation; avoid negative.
Quarter-Kelly (25%) is the common safe default. Full Kelly is mathematically optimal but volatile.
Model probability: 19.7% · fair market: 13.3% · edge: +47.9%
Model probabilities Olympique Lyonnais 58.3% · Draw 22.0% · AJ Auxerre 19.7% Why Olympique Lyonnais is favoured Olympique Lyonnais's Elo (1597) is +143 over AJ Auxerre (1453); high confidence. Recent: Olympique Lyonnais 1W-1D-0L · AJ Auxerre 0W-0D-2L (last 2). Goal differential per game: Olympique Lyonnais +1.0, AJ Auxerre -2.5. Form depth Olympique Lyonnais: avg 1.5 goals scored, 0.5 conceded per game (0/1 at home, 1/1 on the road). AJ Auxerre: avg 1.5 goals scored, 4.0 conceded per game (0/1 at home, 0/1 on the road). AJ Auxerre on a 2-match losing streak. Head-to-head 3 prior meetings — Olympique Lyonnais 0W, 3D, AJ Auxerre 0W (draws are common). Last 3: 3-2, 0-0, 1-3. Calibration Raw model: 67.0% home · Platt-calibrated down to 58.3% based on historical reliability across this sport's graded matches. +EV vs market away: model 19.7% vs market 13.3% (edge +47.9%, suggested Kelly +6.8% of bankroll). draw: model 22.0% vs market 19.7% (edge +11.6%, suggested Kelly +2.0% of bankroll).
No-vig implied probabilities from recorded snapshots. Line moving toward one side means the market is re-pricing in that team's favour — compare to the model to see who's ahead.