Tuesday produced a solid day for the model across 107 races. The first pick won 26 times — a strike rate of 24.3 percent — which is comfortably above the long-run baseline and reflects the predictions doing their job in a competitive field. First picks were placed 58 times in total, meaning the top-rated dog finished first or second in just over half of all races.
The broader picture is even more encouraging. Looking at the top three selections across the card, at least one of them won in 75 of the 107 races — a 70.1 percent hit rate. That means in seven out of ten races yesterday, the model had the winner somewhere within its top three. All three placed in 100 of the 107 races, a coverage figure that reflects genuine predictive depth rather than simply landing the occasional favourite.
This is the kind of day where the model earns its keep: not a spectacular afternoon of long-shot winners, but a consistent, disciplined performance that would have rewarded any structured betting approach. The 24.3 percent win rate for the top pick is meaningful at scale — across a 107-race card, that translates to winners landing at a pace that should generate positive returns on flat staking at SP.
No complaints from Tuesday. The data behaved itself, the model found the winners at a solid clip, and the framework held up on a full card. Today brings another full programme and the model will be working through the data as usual.