Masaki Kouda asked about use of Integrat for asteroid occultations.
If you use Integrat to get an asteroid orbit for the current date, and
use UCAC-2 stars for the star coordinates, you should get a very good
result. It will not be quite as good as the IOTA result, however.

The problem is that the MPC orbits are based on all available
observations, and (to my knowledge) there is no "weighting" of
observations. The IOTA occultation predictions, for example, use
only highly accurate observations from Tycho-2 or UCAC-based
astrometry (plus some from sources such as Hipparcos and the CMC).

Folks such as Steve Preston and Jan Manek take that data, generate
a "fresh" orbit with some weighting of observations, then combine that
with UCAC-2 or Tycho-2 based star coordinates to get a really good
prediction. I believe the situation has basically gone from "this
prediction is a wild guess" (a few years ago) to "this prediction is
probably accurate" (now).

This shouldn't discourage you from running MPC elements through
Integrat to get a "decent, but not exact" result. I haven't done a
lot of this... I'd be curious to see how the results compare to the
IOTA predictions.

-- Bill