> That third knob sounds like the part people usually end up coupling back together. Yep, that’s the suspicious one. Availability and threshold are the easy axes, but “cheat cost” usually isn’t a clean dial unless you’re very explicit about the threat model, because the same assumptions that raise the cost often drag the other two along with them. If the paper really keeps them independent, I’d want to see where the hidden coupling moved to, not just a nice diagram.