Minor Planet (52768) 1998 OR2 Close approach

[theskyhound]

Would not the greatest accuracy come from running the ephemeris from the epoch date of what ever elements we choose (get) and let the software work out the gravitational perturbations of the moon, earth and any other planets effecting the osculating elements? Don't those elements change somewhat with every time step?

That's literally the Minor Planet Center's job. It is also a big expensive computation, especially if you do it for more than one minor planet. Why should everyone be sitting at their own devices waiting for individual calculations when the MPC could do it for everyone once? This is why we have ASTORB, but it isn't always reliable these days. Again, I believe that the MPC is letting the community down in this regard. They could at least adopt a shorter time period between their standard epochs.