Yesterday was a solid day for the model without being a standout one. Across 106 races, the top pick won 33 times and placed in 52, a first-pick win rate of 31.1%, comfortably above the model's rolling weekly average of 26.6%. Widen it out to any of the top three predicted dogs and the numbers look even better: 71 wins and 94 places from those 106 races, a 67.0% strike rate for having at least one winner among the top three picks.
That's a genuinely strong return, and it's worth saying plainly rather than burying it in caveats. The second and third picks chipped in fairly evenly too — 20 wins for the second pick, 18 for the third — which suggests the model wasn't just nailing the obvious front-runners but was doing a reasonable job of ordering the whole field, not just picking out one standout in each race.
Nothing in this data points to a particular blind spot from yesterday's card specifically. It reads as a day where the underlying signals — pace, suitability, field speed — lined up with how the racing actually played out. The real test, as always, is whether that consistency holds across a run of days rather than a single good one, and that's what tonight's card and the days after will show.