Full Vector building has strange eta residual

[kmcdermo]

As pointed in previous PRs, FV building has a strange shape for the eta residual between CMSSW and mkFit backward fit tracks (to the PCA):

• all matched tracks, pT > 0:
  https://kmcdermo.web.cern.ch/kmcdermo/pr126/CMSSWVAL/fit/diffs/SNB_CMSSW_TTbar_PU70_allmatch_deta_fit_pt0.0_CMSSWVAL.png
• all matched tracks, pT > 0.9: https://kmcdermo.web.cern.ch/kmcdermo/pr126/CMSSWVAL/fit/diffs/SNB_CMSSW_TTbar_PU70_allmatch_deta_fit_pt0.9_CMSSWVAL.png
• all matched tracks, pT > 2: https://kmcdermo.web.cern.ch/kmcdermo/pr126/CMSSWVAL/fit/diffs/SNB_CMSSW_TTbar_PU70_allmatch_deta_fit_pt2.0_CMSSWVAL.png

Could be related to 1/pt, which also shows some tail compared to BH or CE:
https://kmcdermo.web.cern.ch/kmcdermo/pr126/CMSSWVAL/fit/diffs/SNB_CMSSW_TTbar_PU70_allmatch_dinvpt_fit_pt0.9_CMSSWVAL.png

Needs to be understood... perhaps the hit assignment or track labeling is messed up?

[kmcdermo]

For sake of simplicity, I am copying the plots directly. As a reminder: "all matched tracks" means all mkFit tracks that are matched to a CMSSW tracks, including possible duplicates.

all matched tracks, diff eta, pT > 0

all matched tracks, diff eta, pT > 0.9

all matched tracks, diff eta, pT > 2

all matched tracks, diff 1/pT

[kmcdermo]

Same comment as other FV problems. will address later

[dan131riley]

By going through the SMatrix Track class, the standard track finders filter out the -1, -3 and -4 fake hits (standard miss, outside and edge detector misses) due to this commit:

f36425c#diff-48dc3f60452a14992adbb68ea6e65a8e

Full vector isn't filtering those out, so it runs out of space to store hits more often than the other methods, especially in the transition regions where there are lots of detector element misses. This doesn't affect the build parameters, but means the backwards fit starts with a large gap from the last to the penultimate hit.

[dan131riley]

I believe this also explains why #128 is specific to hit-based matching. The difference between CMSSW and simtrack hit-based matching is likely the difference in match thresholds.

[dan131riley]

This is not resolved by #134 so it is still a mystery. I also mis-identified the relevant change in CE, I should have linked to

1365812#diff-1164fe0ca653386ef41ffe6b49535295R63

[kmcdermo]

Closing this issue for now, since we are moving away from validation FV. Will re-open if we decide to come back to FV.