🔱 POSEIDON

Probabilities published before kick-off · verifiable track record · public dataset

Bundesliga Predictions

Probabilities for Bundesliga (Germany) matches, computed by a Poisson + Dixon-Coles model and calibrated on real results — published and frozen before kick-off, verifiable after.

5 complete predictions free every day · no account · no ads · zero bookmaker links

The model has analysed 9 matches from Bundesliga in the next 7 days, the first on 23 October 2026.

Upcoming Bundesliga fixtures

WhenMatch1 · X · 2 Over 1.5Over 2.5BTTSMost likely market
23.10 21:30 VfB Stuttgart – Borussia Monchengladbach 🔒 this match's probabilities are for members membership
24.10 16:30 RB Leipzig – SV Elversberg 🔒 this match's probabilities are for members membership
24.10 16:30 FC Augsburg – Union Berlin 🔒 this match's probabilities are for members membership
24.10 16:30 FSV Mainz 05 – Werder Bremen 🔒 this match's probabilities are for members membership
24.10 16:30 1. FC Köln – FC Schalke 04 🔒 this match's probabilities are for members membership
24.10 16:30 SC Paderborn 07 – Hamburger SV 🔒 this match's probabilities are for members membership
24.10 19:30 Borussia Dortmund – Eintracht Frankfurt 🔒 this match's probabilities are for members membership
25.10 16:30 1899 Hoffenheim – Bayer Leverkusen 🔒 this match's probabilities are for members membership
25.10 18:30 SC Freiburg – Bayern Munich 🔒 this match's probabilities are for members membership

Times are Romania time (CET+1). Probabilities are the empirically calibrated ones, not the raw model output. The last column only says which of the tracked markets has the highest calibrated probability — a description of the model's output, not a recommendation; a high percentage is still a probability, not a certainty. The model computes every match in the list; 5 a day are published free, the rest are for members.

How well the model knows this league

On the calibration backtest (January–May 2026, ratings frozen on 31 December 2025) the model saw 173 matches from Bundesliga. It expected 4.5% fewer goals than were actually scored — exactly the deviation the per-league calibration corrects. ✓ calibrated

That deviation is within the ±10% we require to call the league's goal profile separately validated, not just covered by the global calibration.

The full method, with what we correct and what we don't, is on the methodology page.

What came true so far in Bundesliga

30 resolved predictions: 21 hit, 9 missed — 70.0%, with a statistical floor (Wilson 95%) of 52.1%. The sample is still small; we publish it because it is real, not because it is conclusive.

MarketNWINLOSS
Over 1.518135
Over 2.51174
1 (Home)110
Every day we publish 5 complete predictions for free, plus the entire track record — misses included. The rest of the day's predictions, the Discord delivery and the written analyses are part of the membership. Other leagues: all league pages. Această pagină există și în română.

Empirically calibrated probabilities · informational · not betting advice · 18+