ROOT v6.26/08, pullHist shows wrong distribution

avilla · November 18, 2022, 3:22pm

Hi @jonas,
thanks for the tip, I created an issue which can be followed at

ROOT v6.26/08 plots wrong pull distribution with RooPlot::pullHist

opened 11:07AM - 18 Nov 22 UTC

bug

- [x] Checked for duplicates  ### Describe the bug With the recent version of ROOT (v6.26/08), I found out that when plotting the pull distribution after a fit with pullHist, the wrong distribution is shown with respect to the one produced with ROOT v6.26/06, for example ### Expected behavior When the fit converges, a gaussian-like pull distribution is expected, while with ROOT v6.26/08 the histogram bars are unreasonably high ### To Reproduce You can find the needed code and data to reproduce the bug at the following link: https://cernbox.cern.ch/s/hnMmXVjUtbwlZJN Simply run `bug_reproducer.py` with ROOT v6.26/06 and v6.26/08 and see the difference in the plot produced ### Setup Tested with ROOT setup from `lb-conda`

Cheers,
Andrea