If you see a RuntimeWarning about eccentric_anomaly_from_mean_anomaly(), this likely just means the iterative method we use to solve for the eccentric anomaly is getting stuck in a limit cycle. This can happen when your data span a huge range in days. But don’t worry too much about this – for most extreme cases (e.g., when the data span years) the variations are below $$10^{-11}$$. If you see this warning, you can just increase the default tolerance from $$10^{-13}$$ to $$10^{-11}$$ to remove the warning. From the high-level interface (e.g., from JokerParams), pass the keyword argument anomaly_tol=1E-11 to the class on initialization to change the tolerance.