Hanno Rein

Thanks for the details, @KatKiker! I'll try to look into it and see if this is expected behaviour (which I suspect) or something is going wrong...

I can confirm that this is due to a close encounter. The initial conditions which work, get only within 150km of the Earth's centre (note even that is within the...

I would do it at the end of each timestep. One can use the instantaneous positions and velocities and the assumption that particles travel mostly along straight lines to check...

@bhanson10 IAS15 is an adaptive method. You should not adjust the timestep manually. There is currently no way to get the "exact" epoch where a collisions occurs. This would be...

I'm not sure I completely follow! IAS15 is adaptive to begin with. We tested the timestepping algorithm extensively, see https://arxiv.org/abs/2401.02849. So you should not need to do it yourself. Also,...

I see. You could use `reb_simulation_step(r)` which does one timestep of whatever stepsize IAS15 thinks is appropriate.

It looks like you are using a new version of rebound but and old version of assist.

Fair enough. I don't think there's any harm done by adding a few `#ifdef`s. Doing anything more complicated seems like an overkill. (Probably not what you want, but you won't...

I'll let Matt comment on this, just one small thing: note that an inverse rotation is performed [a few lines further down](https://github.com/matthewholman/assist/blob/main/src/forces.c#L534C9-L534C14).

Maybe I'm not up-to-date with all the python packaging stuff, but I'd be cautious to install large data files as python packages. (You already mention that you're waiting for a...