A large fraction of failure in MC truth-matching when using Retina Clusters in Velo tracking reconstruction

As discussed in mattermost channel

• The reconstructed kinematics when using VeloSP or Retina clusters match each other very well
• but quite large fraction of reconstructed tracks can not be matched when using Retina Clusters

2023 MC sample used: root://x509up_u128460@eoslhcb.cern.ch//eos/lhcb/grid/prod/lhcb/MC/Dev/DIGI/00194513/0000/00194513_00000005_1.digi Also checked with 2022 minbias MC: root://eoslhcb.cern.ch//eos/lhcb/grid/prod/lhcb/MC/Upgrade/XDIGI/00171960/0000/00171960_00006077_1.xdigi

See more plots in attachment. truthmatch_20231007.pdf

The PrTrackRecoDumper algorithm was used to dump the reconstructed info, with this option: /afs/cern.ch/user/p/peilian/eoslhcb/public/forGiovanni/hlt2_trackrecodumper.py