Trying to use advise function for optimizing, means I modelled some conditions for entering a trade and want to know if a machine learning algo will find some additional suggestions to do so. There is one problem: scaling values within the advise call after an if condition results in a '041 Inconsistent series calls'. Is there any way around? Or do I have to use algos that work with unscaled values?

Thanks, Sphin

Your computer does certainly not want you to use unscaled values. It rather wants you to scale it in the right way, i.e. not inside some if conditions.