We discuss four important problems in automata theory each of which has three natural variants: a deterministic one, a nondeterministic one, and as an in-between case a nondeterministic one with multiplicities. Our goal is to support a general view that the third variant is much closer to the first one than to the second one.