Big, bright fast mover (7482) 1994 PC1

[theskyhound]

I'm thinking ST4 uses the element set closest to the date of requested ephemeris. That's why the positions in my table were much better for the daily and osculating vs the default in the bottom line (standard MPC import).

Yes, SkyTools keeps multiple sets of osculating elements and chooses the one that is best for the date that a position is being calculated for. But it is very important to remember that the epoch that the osculating elements were calculated for is just one factor in how accurate the elements will be. These close approachers are faint, so they can't be observed except when they come close to the earth. As a result, their orbits are not well known due to inadequate astrometric observations.

HORIZONS may appear to be more accurate during a close pass, but the real determination of accuracy is calculating the orbit using new observations as they become available, and new ones tend to come in just before the pass, so the accuracy of the elements gets better and better as that time comes closer. Always update your orbital elements just prior to the pass.