But why does everyone recommend they be tied to ground if it does nothing?
Any why does the signal wire RF penetrate the cable shield (I assume this is what you are saying) when the shield is isolated from a box, and not when it is connected?
I'm not sure how much shielding theory and EMC this forum will accept.
BUT, A shield requires no connection to earth(dirt) to function. There are plenty of shields inside your cell phone that work very well and see no wires to earth.
Same with a handheld VHF.
In regards a wire having to penetrate a shield:
Let's say you have a perfect shield, but, of course you need to power the device and get inputs and outputs from this shielded device.
One cannot violate the shield by running a bare, unshielded wire into the shield. WHat you would have is an antenna, picking up energy on one side of the shield wall, and transferring it the to other side of the shield wall. And, that effect is bilateral (or commutative??). So, what is done is one of several things.
1. A filter is fitted at the shield wall that suppresses unwanted energy in the wire of interest. Lumped or discrete inductive and capacitve elements typically, or,
2. The wire is shielded itself, and will need to connect to the wall of the product shield. This, in effect, extends the product shield to the wire.
In no case does a "wire" connected to the shield WALL going to earth, or any other place, make a
shield work better.
Now, there are occasions where a bond wire to the shield makes a
product work better. It won't be due to shielding improvements, but more likely an antenna counterpoise, or other effect.
HF/SSB, and LoranA/C, Omega, etc are low frequency devices, and the associated marine antennas can benefit from a counterpoise. Which COULD be done by "grounding" the case of the receiver, or tuner.
One of the "other effects" is a static bleed, due to antenna charge buildup. Those you want to bleed to earth, via that thumbscrew/wingnut. That can reduce corona, st elmo's etc, either will cause receiver noise.