As George mentioned, my experience is that the primary reason for the difference between Occult 4 output and Occult-watcher is the fit used. The JPL fits include Gaia DR3 data on 150K+ asteroids and the MPC fit does not. I still make a first past using the MPC data and then refine with the JPL fit on a subset of the events. The reason is that the JPL fit calls Horizons which takes a longer time, and it is faster to run using the MPC data. The JPL data so far does not include the most recent Gaia data which extends to 2020, so for the moment, the Gaia data are from 2014-2016.
If Occult-4 is crashing, maybe I can help here. We got it to work, so y'all should be able to. Can you post a screenshot with the exact error message? It may be in the orbital parameters file. Knowing which rock it crashes on would be helpful. Before doing that, update to the latest MPCorb file and latest Occult-4 version.
The reason for uploading Astorb and Ast-DyS-2 first is that they include fit error data in their file and MPC does not, so the errors are first calculated using the other data-sets. Of the 3, MPC seems to have the largest data-set.
I still get some fits out of Occult-4 that change dramatically when uploaded to OW and they do a fit, and I haven't tracked that down as yet, but it is only a couple of percent of events that do that now, so I just live with it. Most of the events generated agree with what gets posted on OW.
