I think it would be really useful to be able to override the RooAbsPdf default error level using a name argument to the RooAbsPdf::fitTo() method e.g.
As far as I can see there is no reason why this could not be implemented and it would be very useful in conjunction with RooMCStudy for which at the moment (as far as I can see), one has to re-do the entire (potentially computationally expensive) fit if one wishes to evaluate e.g. a confidence interval at an error level other than the default (c.f. RooStats::UpperLimitMCSModule). I may be mistaken in my understanding and would be happy to hear of a way to get e.g. a 95% confidence interval for each iteration of RooMCStudy without refitting every time (or deriving a new PDF class with a different default error level!).
Cheers,
Hugh