This system has a clear logic and the recent numbers back it up. The premise is straightforward: take the composite top pick at Doncaster, but filter out four specific grades -- A2, B5, D2, and D3. Those grades tend to produce higher variance results, either because the fields are tightly packed with little form differentiation or because the dogs involved have inconsistent profiles. Remove them, and the remaining pool of races gives the top composite pick a cleaner run at predictability.
The results this week: 13 qualifying runs, 7 winners (53.8% strike rate), 8 placed (61.5%). The win ROI sits at +53.4%. That is an exceptional short-term figure.
Honesty matters here. Thirteen runs is a small sample, and a 53% strike rate over any dozen-plus bets can appear by chance. What makes this worth monitoring beyond the numbers is the structural argument underneath it. Grade-exclusion systems do not just filter noise by accident -- they reflect a genuine observation that certain race types reward form assessment more reliably than others. If that observation holds over a larger sample, the edge should persist.
No qualifying picks from the system today, so it is currently waiting for the right conditions at Doncaster. When it does fire, the recent form suggests it is worth paying attention. You can explore this system and set up your own filter combinations on the [Systems page](/systems).