During UML fit of some distribution i find the fitter gives me yield which is 0.2 +/- 25.0
(the error is 100 times larger)…the EDM is ~ 10^-5 or so… not sure if it is possible for roofit
to give me this large error and fitter still converge? Not sure, if it makes sense ?
it looks like a failed fit. Have a look at the fit output. Minuit tells you what it thinks of the fit. It looks something like what I pasted below. Already the first four lines are quite instructive …
MIGRAD MINIMIZATION HAS CONVERGED.
MIGRAD WILL VERIFY CONVERGENCE AND ERROR MATRIX.
COVARIANCE MATRIX CALCULATED SUCCESSFULLY
FCN=4.2579e+06 FROM MIGRAD STATUS=CONVERGED 37 CALLS 38 TOTAL
EDM=4.50916e-07 STRATEGY= 1 ERROR MATRIX ACCURATE
EXT PARAMETER STEP FIRST
NO. NAME VALUE ERROR SIZE DERIVATIVE
1 mean 9.99446e-01 5.77560e-04 8.27114e-05 -8.06984e+00
2 sigma 1.00036e+00 4.08396e-04 2.04434e-04 -3.35168e+00
ERR DEF= 0.5
EXTERNAL ERROR MATRIX. NDIM= 25 NPAR= 2 ERR DEF=0.5
3.336e-07 1.370e-10
1.370e-10 1.668e-07
...