@Domonkos Most of the nested state machines I’ve seen are based on a complex functional breakdown. Redrawing the nesting on a single-level state diagram ends up looking like spaghetti and meatballs. You usually don’t have to redraw much to make the point that you have a very complex mess. I haven’t ever seen an example put online.
In Executable UML, we manage complexity in the class model, so the complexity doesn’t get pushed down into the state machines and action language (process models). This exposes dependencies and constraints much better.