See rejected points in a fit

Maybe this … if you use the chi2 method, bins with zero-errors (a.k.a. “empty bins”) will usually be excluded.
See TH1::Fit and [url]Empty bins treatment when fitting
Another potential source of problems is the “R” fit options -> if the range specified in the (fit) function range is too narrow, some histogram bins will be excluded.

(And yes, I know I didn’t answer your original question.)