Fix identical track check for LHCb ID/VELO ID overlap table

In a discussion on !4463 (merged) it was noted that a large change was visible in the VELO overlap relation table due to a change in the persistency. This is related to a pointer check on the tracks, which crept in via another recent MR. The outcome of the current test in DaVinci for this functionality is not correct, as it does not match particles which have the same tracks.