I wonder now more about the science of impact.
Thinkin' out loud...with a piston stop and wrench with torque slowly and gently applied and increased, the torque applied to the nut is transferred through all the parts all the way to the piston stop. They are likely not excessive under normal circumstances. When the nut breaks free, the forces go to zero on all the parts like the crank and piston and connecting rod and wristpin, etc. Duh.
With the impact, a jolt is applied to the nut. The flywheel has significant mass and rotational inertia, as does the crank, significantly more than the nut. The lobes on the crank, and the flywheel itself, are there for rotational inertia. A quick force spike to the nut breaks the nut free before the flywheel and crank can begin to rotate. Because of their inertia, from the crank and flywheel's perspective, barely anything happened before the nut flew off. Near zero forces and torques are therefore applied through the piston, con rod, and wristpin, etc., assuming the nut breaks free relatively easily.
In the case where the flywheel is held still, even less is transferred to the other parts, but I guess one could break a flywheel fan blade that way. However, the radius from the outside of the flywheel to the crank center may be, say, 3 inches while the radius between nut threads and crank is say, 0.25 inches, Since the outside of the flywheel is further away, for a given torque, it takes less force to hold it in place against the nut torque. The fan blade is unlikely to break, especially if multiple blades are held with a rag, etc.
With the impact, a jolt is applied. We really don't know how hard that jolt is, which is probably the biggest risk. A jolt in the wrong direction may be even worse. Application of steady torque with a wrench can be easily measured.
I like impact.
Witchcraft?