loading…
Rating is 0–100 against every runner since 2019, so it compares across races;
Win % is the calibrated
probability and sums to 100% across each race. Fair is 1/probability — the price
you would need. Fcast is the racecard's forecast price and Diff
compares the two.
Today's price is never an input to the model, so Diff is a genuine disagreement with the forecast
rather than a restatement of it. It is not a value signal: tested against real exchange prices, the
model's disagreements did not produce a profit at any price point.
Click any column heading for the detail behind it.