Yesterday's card covered 169 races, and the model returned a first-pick win rate of 32% — 54 winners from 169 top selections. That sits well above the 7-day average of 27%, making it the best-performing day of the past week.
The place figures backed up the win numbers: 77 of the 169 top picks finished first or second, a 45.6% place rate. Across a full card, that represents consistent form identification rather than a cluster of short-priced winners distorting the headline figure.
Expand the net to the top three predictions and the numbers get particularly strong. One of the model's top three picks won 119 of the 169 races — 70.4% of the card. In 152 races out of 169, at least one of the top three selections finished first or second. That is 89.9% coverage of the frame from three selections per race.
The benchmark matters: a randomly selected dog in a six-runner field wins 16.7% of the time. At 32% on first picks, the model is generating roughly double that rate. One good day is not a trend, but it adds to a 7-day stretch that has averaged 27% — consistently above baseline.
Full results from yesterday are available in the [historic results section](/historic-results?date=2026-06-17).