Wrong fit parameter

Hi Everybody.

I’m fitting a histogram with a “gaus(0)+pol2(3)” and the third parameter (p2), which is the sigma of the gaussian, came with a negative signal,

I attach a rootfile (fit_rho.root) which contains 2 histos: “hrho_fit”, a histo already fit and “hrho” to be fited by a rootscript “rho_fit.cc”

Can anyone tell me if I’m doing something wrong? Or if it is a bug?

Cheers, Luís.
fit_rho.root (8.48 KB)
fit_rho.C (333 Bytes)


Add the statement
to for ce a positive value.
ROOT fitting forces the positive value only for bare gaussians,
not for expressions including gaussians, although this improvement could be done.