Hi,
I have a RooDataHist object which is fitted with my PDF function. I am satisfied with the fit values, but I know that the errors are wrong. To calculate them “right” I want to use RooChi2Var object, but only for the errors calculation, not for a new fit. But unfortunately the RooChi2Var starts a new fit, which I don’t want (because chi2-fit does not converge in my case).
Any ideas how to do that?
Thanks in advance,
Alexander.
RooDataHist &data=....;
RooAbsPdf &pdf=....; // the pdf after fit, values are good, errors are bad
RooChi2Var chi2("chi2","chi2",pdf,data,DataError(RooAbsData::SumW2));
RooMinuit m(chi2);
m.hesse(); // NEW FIT! I DON'T WANT IT!
RooFitResult *res = m.minos(); // NEW FIT! I DON'T WANT IT!
[/code]