search

cms-sw / cmssw

CMS Offline Software
http://cms-sw.github.io/
Apache License 2.0
1.07k stars 4.29k forks source link

Unphysical IsoTrack pt #45276

Open vlimant opened 3 months ago

vlimant commented 3 months ago

Mirroring an issue opened under xpog gitlab https://gitlab.cern.ch/cms-nanoAOD/xpog-coordination/-/issues/40 so that this can be followed up on here.

There are events with unphysical IsoTrack_pt, e.g. in WF 2500.312 (ttbar UL18):

image

In that WF, all the 11 tracks with pt>6500 are flagged as isPFcand and isFromLostTrack.

cmsbuild commented 3 months ago

cms-bot internal usage

cmsbuild commented 3 months ago

A new Issue was created by @vlimant.

@antoniovilela, @Dr15Jones, @smuzaffar, @makortel, @rappoccio, @sextonkennedy can you please review it and eventually sign/assign? Thanks.

cms-bot commands are listed here

VinInn commented 3 months ago

Is anywhere supposed to be a clamp of pt? Surely not at track level, 1/p(t) can be zero for what the track fit is concerned.

slava77 commented 3 months ago

ah, this is just about large values of track pt. I misread as if there were NaNs and infinities when in gitlab 40 I suggested to open an issue here.

Now that I read this more correctly, I agree with Vincenzo. This is a non-issue

vlimant commented 3 months ago

You're going too fast for me ; are you saying that 40TeV track are physical ?

slava77 commented 3 months ago

You're going too fast for me ; are you saying that 40TeV track are physical ?

yes. well, at least within uncertainties. What does the ptError() return for these tracks? If it comes out below 20% or so, I'd be motivated to follow up.

Dr15Jones commented 3 months ago

assign reconstruction

cmsbuild commented 3 months ago

New categories assigned: reconstruction

@jfernan2,@mandrenguyen you have been requested to review this Pull request/Issue and eventually sign? Thanks

slava77 commented 3 months ago

What does the ptError() return for these tracks? If it comes out below 20% or so, I'd be motivated to follow up.

actually, compare qoverpError() with charge()/p() to stay on a Gaussian side of interpretation of uncertainties. A track with abs(qoverpError()*p())<0.2 and pt of 40TeV would be worth debugging.