Spolytics Research Note: September 6, 2026 Model Outputs
On September 6, 2026, the model evaluated 30 games across MLB (15 games), La Liga (4 games), Serie A (4 games), Ligue 1 (3 games), EPL (2 games), and Bundesliga (2 games). The system derived output estimates by evaluating expected goal and run distributions across each scheduled fixture.
Probabilities computed that day
- 90.0%
FC Augsburg at Eintracht Frankfurt
분데스리가 · Both 총 2골 이상 · O 1.5
- 88.0%
Barcelona at Valencia
라리가 · Barcelona 1골 이상 득점 · 1+ Goals
- 85.0%
Stade Rennais at Angers
리그 1 · Both 총 2골 이상 · O 1.5
- 84.0%
Paris FC at Marseille
리그 1 · Marseille 1골 이상 득점 · 1+ Goals
- 83.0%
Manchester United at Everton
EPL · Manchester United 1골 이상 득점 · 1+ Goals
- 83.0%
Mainz at Hamburg SV
분데스리가 · Mainz 1골 이상 득점 · 1+ Goals
Highest Computed Probabilities
In the Bundesliga fixture between FC Augsburg and Eintracht Frankfurt, the model calculated a 90.0% probability for Over 1.5 goals based on a combined expected goal value of 3.93. In La Liga, Barcelona scoring 1+ goals at Valencia reached an 88.0% probability from an expected team goal figure of 2.12. Stade Rennais at Angers in Ligue 1 registered an 85.0% probability for Over 1.5 total goals with a combined expected goal figure of 3.36. Also in Ligue 1, Marseille scoring 1+ goals against Paris FC logged an 84.0% probability with 1.82 expected team goals. Finally, Manchester United scoring 1+ goals at Everton in the EPL and Mainz scoring 1+ goals at Hamburg SV in the Bundesliga both generated an 83.0% probability, with expected team goals of 1.76 and 1.75 respectively.
Methodology and Scored Performance
The probability framework utilizes public statistics APIs to establish base rates, applying capped adjustments before performing daily refitting. Across 21 logged days of daily refitting, the model has processed 34152 graded samples. Cumulative historical performance tracking reflects a Brier score of 0.201 alongside a hit rate of 69.1%.
These figures represent statistical probability estimates with error rather than guaranteed predictions. All outputs serve as observational metrics calculated from historical base rates and distributional modeling distributions.