Hi Claudio,

Getting asteroid occultations right is a pretty tough job.
You can do it in Guide, but you won't get there by using
elements from MPC or from the Guide CD.

Jan Manek (and Steve Preston and others doing this sort of
work) get accurate results by computing their own orbits,
using only high-precision astrometry. MPC elements, and
those on the Guide CD (which came via Lowell Observatory's
ASTORB database), use astrometry from a host of sources;
some are far less accurate than others, and I don't think
either system does any weighting of observations (i.e.,
"these data should be counted much more strongly in the
solution than these other data.")

There are some hints on how to do all this at

http://www.projectpluto.com/ast_occu.htm#do_it_yourself

The biggest problem is getting accurate orbital elements
for an epoch close to that of the occultation. If you have
the raw astrometry, you can feed it through Find_Orb and
get a good orbit and use that. It does take some patience,
but some people have done it. Steve Preston gave it a try
a while back, in order to get a check on computations done
with Occult. There were some small differences between
Guide's results and Occult's (a few kilometers). That was
a bit maddening, since we could see no reason for there to
be any difference at all; but we also can't really tell
which program is actually in error, because the accuracy
of either prediction is more like dozens or hundreds of
kilometers.

-- Bill