Statistics
Aggregated AI prediction performance.
How success is measured
Winner
Counted as successful when the AI's predicted winner finishes 1st. Any other finishing position (or a non-runner) counts as missed.
Place (each-way)
Modelled on a standard each-way bet. Counted as successful when either the predicted winner or the main danger finishes:
- 1st or 2nd in races with 8 runners or fewer
- 1st, 2nd or 3rd in races with more than 8 runners
Otherwise counted as missed.
Top pick placed
A focused view of the predicted winner alone. Counted as successful when the predicted winner finishes 1st, 2nd, or 3rd (top 3 only in races with more than 8 runners; top 2 otherwise). Unlike the each-way metric, the main danger is not considered.
Only races with confirmed official results are graded. Pending races are excluded from the totals above. Informational only — not betting advice.
Winner predictions
Place predictions (each-way)
Successful when the predicted winner or main danger places (top 2 in fields of 8 or fewer; top 3 in larger fields).
Top pick placed
Predicted winner finished in the place positions (top 2 in fields of 8 or fewer; top 3 in larger fields).
Accuracy — last 30 days
No data.
By date
| Date | W ✓ | W ✗ | Winner % | P ✓ | P ✗ | Place % | TP ✓ | TP ✗ | Top pick % |
|---|
Experimental model
We're testing a new 12-factor weighted scoring model that pulls in extra signals like jockey/trainer 14-day form, course bias, pace style and surface preferences. These are early results, kept separate from our main numbers above so you can judge each system on its own merits.
Based on 0 races analysed in the last 90 days.
Winner predictions
Place predictions (each-way)
Successful when the model's predicted winner or main danger places (top 2 in fields of 8 or fewer; top 3 in larger fields).
Top pick placed
The model's predicted winner finished in the place positions (top 2 in fields of 8 or fewer; top 3 in larger fields).