Straight from the author:
I don’t really have much more to contribute. “Safe” is a relative term; I agree you just have to try different values and see if you believe the results. The “default” value was based off the benchmarks in my paper, to see what gave the best results.
I don’t remember how the mobile residues are determined exactly. I think it’s all residues within X Angstroms of the ligand, plus ~2 residues up- and down-stream of those in sequence order. Someone would have to dig in the code to see what X is.
A note from me: “X” appears to be residues having a backbone atom within 7 angstroms, if you look at the code. I could be wrong.