Thursday produced a workmanlike day across 130 races, with the model hitting 31 first-pick winners for a strike rate of 23.8%. That sits just below the long-run benchmark of around 25%, but over a 130-race sample any single day will fluctuate around that figure — Thursday was not a day that flatters, but it was not one to worry about either.
The more telling number, as always, is the coverage figure. Any of the top three predicted dogs winning 80 times from 130 races represents a 61.5% coverage rate — meaning nearly two in every three races were resolved by a model pick in the top three. That kind of coverage across a high-volume card is what systematic selection approaches depend on, and Thursday delivered it comfortably.
First-pick place performance was also encouraging, with 62 of 130 selections finishing in the frame — a place rate of 47.7%. For those running place-based systems, that is a productive baseline to work from.
No single day tells you much on its own. What matters is the pattern over weeks and months, and Thursday's numbers sit within the expected range for a model operating at its current calibration. The top-three coverage being as strong as it was suggests the model is identifying the right races and the right dogs even when the winner comes from the second or third selection rather than the first.
Full results and grade-by-grade breakdown are available in the historic results section.