As far as i concerned, the fNpx
variable affect the TF1::GetRandom
(and its overload). I am wondering if that fNpx
also affect the fit results on histograms and graph (afterall they are binned data)?
so I generated a pseudodata and fit with functions of (exaclty) the same form but different Npx, and I ended up with exactly the same fit results (parameters, their errors and chi2).
is it the way that TF1 works (Npx setting doesnt have effect on fitting) ?
ROOT Version: 6.24.00 (conda-forge)
Platform: ubuntu 20.04 (wsl)
Compiler: gcc9