Uncertainty in fit parameters vs statistical uncertainity?

Hi all,

I fit a histogram with an exponential function; ROOT gives the Slope and the ERROR in the slope.

My query is, is the error in the slope already including the statistical uncertainty 1/\sqrt(n), n is the histogram integral or is it purely coming from the (co)variances?

Thanks a lot !

Dear @utkarsh-2000 ,

Thank you for reaching out to the forum! I will let @moneta comment on your query. Meanwhile, did you already have a look around the documentation? If this information is missing or unclear we could add it.



The error in the slope (fit parameter) comes from the fit result covariance. Since the fit includes the statical error of the histogram bins, the covariance includes the statistical uncertainty from the bin content fluctuations.

Best regards


1 Like

Thanks a lot @moneta