Skip to content

Positivity errorbar from errorbar68

Emanuele Roberto Nocera requested to merge positivity-errorbar into master

Created by: wilsonmr

closes #961 (closed)

As discussed in issue, using the central value and standard error with MC pdf sets makes it look like there are negative replicas, when there aren't any.

This still uses the central value as the reference point for the errorbar plot. Potentially we could end up with something like the arclengths where the point is outside of the errorbar.

I didn't change this initially because it's using the central value of the stats class. If we had a hessian PDF then this is fine. I could check if the pdf is MC and use the median instead. Another option would be the remove the central value point. Does anyone have a strong opinion?

In practise I guess we probably rarely use this on hessian PDF sets

Merge request reports