[TowerTalk] "Noisy" ground with a balanced antenna?

[Bill VanAlstyne, W5WVO]

I've read in a couple places lately -- maybe in ads, not sure -- that
creating an "effective RF ground" (whatever that would imply at the
wavelength in question) can reduce received noise in the antenna circuit,
even when an "RF ground" isn't necessary for antenna current return, such as
when driving a correctly-tuned balanced antenna. Is this true? I work a lot
of 6M SSB/CW, and man-made QRN is often the limiting factor on receiving
weak sigs. My 5-el homebrew yagi presents a virtually pure-resistive 50 ohms
load at 50.125, according to my 259B. Wouldn't expect any improvement on the
transmit side. But would my receive noise floor really drop if I put down a
counterpoise under my carpet near the rig? Or outside? If so, why? I don't
see it. Please enlighten me as to how antenna currents developed by noise
sources could be reduced with an "RF ground". Given that the antenna,
feedline, and radio are perfectly matched to each other at 50 ohms, where
would these currents be flowing, and how would an "RF ground" reduce them?

Bill / W5WVO
New Mexico / DM65