Permanent, self-updating scorecard for the BeyondThePrem ML model. Whenever you want to know how often the model’s top call has actually landed this season, this is the URL. Updates automatically as each round’s results flow in.
IMPORTANT — NOT AN EV PAGE
Everything below tracks the ML model’s top-call accuracy — how often the outcome the model rates as most likely actually happens. That is different from EV selections, which are picks where the model’s probability beats the market’s implied probability.
For EV-specific accountability (stake on / return on / ROI), see our model accountability posts.
Season-to-Date — All Leagues Combined
Running Scorecard
CHAMPIONSHIP
67/321 (21%)
LEAGUE ONE
73/300 (24%)
WSL
—
Each cell: correct / total predictions (hit rate). Random baseline for a 3-way market is 33%. A trained model should clear that over a sufficient sample — and we are honest when it does not.
All-League Stats Dashboard
Combined Stats
Prediction Stats (2025)
Championship
| Predictions | 107 |
| Played | 107 |
| Points | 67 / 321 |
| Correct scores | 11 (10.3%) |
| Correct results | 45 (42.1%) |
League One
| Predictions | 100 |
| Played | 100 |
| Points | 73 / 300 |
| Correct scores | 12 (12%) |
| Correct results | 49 (49%) |
Championship — Model Scorecard
Season-to-Date Stats
Prediction Stats (2025)
Championship
| Predictions | 107 |
| Played | 107 |
| Points | 67 / 321 |
| Correct scores | 11 (10.3%) |
| Correct results | 45 (42.1%) |
Accuracy by Matchday
Points by Matchday (2025)
Accuracy by Predicted Outcome
Prediction Breakdown (2025)
By Team
Avg Points by Team (2025)
Common Score Patterns
Most-Predicted Scorelines (2025)
League One — Model Scorecard
Season-to-Date Stats
Prediction Stats (2025)
League One
| Predictions | 100 |
| Played | 100 |
| Points | 73 / 300 |
| Correct scores | 12 (12%) |
| Correct results | 49 (49%) |
Accuracy by Matchday
Points by Matchday (2025)
Accuracy by Predicted Outcome
Prediction Breakdown (2025)
By Team
Avg Points by Team (2025)
Common Score Patterns
Most-Predicted Scorelines (2025)
WSL — Model Scorecard
Season-to-Date Stats
No prediction data found.
Accuracy by Matchday
No matchday prediction data found.
Accuracy by Predicted Outcome
No prediction breakdown data found.
The honest framing
A probability model’s output is not a forecast you can bet on like a weather app. It is an estimate of how likely each outcome is, given the features the model has been trained on. We track these numbers in public because:
- Accountability keeps us honest — we publish the bad rounds as loudly as the good ones
- Small samples can lie — the numbers above only get statistically meaningful as the season accumulates
- The right metric depends on the question — hit rate is easy to read, log loss is more rigorous, and neither answers “is this a good bet”
For the “is it a good bet” question, see our EV-filtered Heinz and Yankee posts and the follow-up accountability write-ups.