I agree at some point optimisation becomes a case of diminishing...

I agree at some point optimisation becomes a case of diminishing returns (and there are good reasons not to over-optimise in the first place) so at that point it is worth simply running it in demo to test correction execution and then go live - so long as the strategy is profitable it will start making money and even if not optimal (i.e. its not making the most possible money you think it can make) continuous improvement can be used to make improvements over time to improve it.

Cheers, Sharks