CPE in HLT: pixel track dxy vs phi large deviations

[slava77]

We observe (since a while) large deviations of mean dxy vs phi for HLT pixel tracks

Data e.g. https://indico.cern.ch/event/1319686/#10-tracking-pog

MC:
https://tinyurl.com/24m77lgd

Both vary in range roughly ±80 µm, compared to ±10µm or less for full tracking.

It is likely that the origin is from the fast CPE, but could be some other feature in (patatrack) HLT pixel tracking.

@AdrianoDee @vjmastra @elusian
you may want to follow this as well