Best of luck for March,
MartinK. I don't know the intricacies of the system, but I'm forever wary of backfitting systems without good reasons, so just some observations that you're obviously free to ignore.
Cutting out heavy ground could make sense - I often remove the extremes of a criteria.
But is there a reason why Class 5s should be less profitable than others? If the extreme low grade (class 6) is profitable and the middle ground (class 3/4) is profitable, what logic is there to say class 5 isn't? Some winners might be around the corner there.
Same question about hurdles. I think it's reasonable for systems to distinguish between codes, but if it works on the flat and works in chases, why wouldn't it work for hurdles?
My final comment is an obvious and general observation: good systems and approaches have losing runs, and there's every chance you're changing it just as your existing February losing system would pick some winners/your new March system would hit a bad run.
As I say I don't know your system so there may be good reason to cut class 5s and hurdles, but hopefully some helpful things to think about.