Dear @jonas, to possibly facilitate the analysis of this problem I attach to this message the full code I am using, as well as the two ROOT files containing the data to be fit and the MC-derived histogram from which I extract the PDF.
To run it, python3 ana_globalFit.py 1 (1 is for “period_1.root”).
As you see, if you comment line 161 and uncomment lines 164 and 165 you will move from a simultaneous fit to a one-by-one fit. The latter works, the first does not.
Thanks again
Andrea
code.tar (33.3 KB)