Mark, you are again making my point (regarding the leak): If there is enough of a leak to be detected at a post-cal, it would also show on the two rides. The two counts would not be in-tolerance, and the measurer would try one more. That, also, would be out of tolerance, so the measurements would have to be discarded. Post-cal would not be necessary, since the measurements were invalid.
Further, if one set an on-site calibration check course, they could verify good measurements. Not sure why anyone would not use an on-site check course, to verify no problems before leaving the course.