Extrapolation of (Roo)FitResults

Dear RooFit experts,

I am doing an Extended likelihood fit, using a PDF

F(x)= N1 * S1(x) + N2* S2(x)

in range of a histogram (I call this range “control region”) . I would like now to extrapolate the Fit Results to the other region (“signal region”), assuming that the shapes S1(x) and S2(x) are known in the full range.

My question is what is the correct way to extract the N_total error in the Signal region .

I have created RooFormulaVar s for the Number of events in the “signal region” as:

RooFormulaVar sN1("sN1","r1*N1",RooArgList(R1,N1));
RooFormulaVar sN2("sN2","r2*N2",RooArgList(R2,N2));
RooFormulaVar ntot("ntotal","sN1+sN2",RooArgList(sN1,sN2));

with r1, r2 the ratios between the control and signal region shapes (which I take as constants here) .

Is the error extracted as :


correct ?
And how can I extract Minos asymmetric errors if exist ?

Many thanks