I think you are correct, but I like freebees, ... Do you have an online demo?
Is the maximal speed of the slingshotted particle something like twice the speed of the big objects? Do they escape or they just form a cloud that gets hotter?
For this subject, it's important that the simulation conserves energy. Most naive numeric simulations don't do that, and even symplectic simulations increase/decrease the total energy slowly.
> Do they escape or they just form a cloud that gets hotter?
They escape.
yeah, having done this before, I am 100% sure that the simulation will be inexact due to numerics, and that the inexactness will be worse around these phenomena, but we are probably talking, gut feeling say, under 20% in most scenarios where you see an escape (I think the observable error is effectively unbounded because you could in theory get two particles within one ULP or even two particles that collide and cause a NaN error)... But in general I don't think that changes the qualitative nature of the phenomenon. There's probably a reasonably easily derived "starting from three particles at rest" where you can see one of them escape from the other two; if not 3 then four.
Is the maximal speed of the slingshotted particle something like twice the speed of the big objects? Do they escape or they just form a cloud that gets hotter?
For this subject, it's important that the simulation conserves energy. Most naive numeric simulations don't do that, and even symplectic simulations increase/decrease the total energy slowly.