Chi-square definition in ROOT

[Wile_E_Coyote]

@moneta

In the case of binned likelihood fit I agree this is confusing, it returns the Neyman chi2 using your fitter function.

This is not the case.
See the numbers in my first post here (“Neyman chi2 = 7.94587” versus “f->GetChisquare()” which returns “8.2006239” ).

It seems this has already been raised a year ago or so (you fixed something but it’s still incorrect):

Getting details of a gaussian fit

@moneta In your example macro, if you set “gStyle->SetOptFit();”, in the “statistics box” of the “Maximum Likelihood Fit” you will see “chi2 / ndf” which is completely different from “2 * FCN”. The value that is displayed is the same as returned by “result2->Chi2()”.