TAHMIN MODELI GUIDE
How to read performance statistics
To understand a performance rate, consider the metric and number of evaluated races as well as the percentage. Ranking a winning horse and correctly selecting every leg of a sixfold measure different outcomes.
What do top pick and top three mean?
The top-pick rate is the proportion of evaluated races won by the report's first-ranked horse. The top-three rate is the proportion where the winner appears among the first three selections.
For example, 6 first-pick winners from 20 races means a 30% top-pick rate. If 12 winners are in the top three, that rate is 60%. These are examples, not the website's current rates.
Is this the sixfold winning rate?
No. Ranking one race winner correctly differs from fitting all six winners into a coupon. Budget, selected horses and the combined outcomes of six races affect coupon performance.
Homepage winning sixfolds and the ranking performance chart should therefore be considered separately. A high rate for one metric does not mean every coupon will win.
How are the model, AGF and Sixth Sense compared?
The model represents performance rankings, AGF the relevant market preferences, and Sixth Sense a separate supported ranking. Each series may have a different number of eligible races.
A race may be excluded from a series if an eligible prediction or required source record is missing. Evaluated and excluded counts should be read separately. A day with no evaluations is not the same as a day with a 0% success rate.
Why do dates and sample sizes matter?
With a small sample, one result can change the percentage substantially. Two correct picks from four races is 50%, but it is based on fewer observations than 50 correct picks from 100 races.
Daily tables and charts show recent changes. Longer periods, different racecourses and race types provide a broader assessment. Past performance does not guarantee future results.
Frequently asked questions
Where can I see current performance rates?
Performance summaries appear in the homepage model cards. Sign in to access the detailed performance screen.
How should a day without data be interpreted?
If a series has no evaluations for a day, its percentage cannot be calculated. This is different from a day when all predictions were incorrect.