Transparency matters, so here is an honest look at how the prediction models have fared over the past seven days. Across 1,261 races the top pick won 22.3 per cent of the time and placed in 42.1 per cent. Widen it to all three top-rated dogs and the picture is healthier: the winner came from our top three in 58.6 per cent of races, and at least one of the three placed a hefty 83.7 per cent of the time. In other words, our shortlist of three is finding the frame in more than four races out of five.
The week was not uniform, though, and the day-to-day swings are worth understanding. The strongest days were the ninth and the seventh of June, with the top pick striking at 30.2 and 29.8 per cent respectively, both well clear of the average. The toughest came on the fifth, when 264 races returned just 15.9 per cent for the first pick, and the eighth at 18.8 per cent. Those lower days share a feature: they were among the busiest, and big-volume days tend to include more low-grade and maiden races where form is patchy and upsets are common. When the cards are thick with hard-to-weigh races, the strike rate naturally dips.
The steadier story is the top-three place rate holding above 80 per cent throughout. That consistency is what makes the predictions useful even on the rough days, because while the single best pick can be foiled by a bad break, the model is reliably putting the right dogs on the shortlist. Plenty working well, then, with low-grade volume the main thing dragging on the headline strike rate. We keep watching it.