Hum interesting but I think based on a false premise that the grid
has the *exactly* the same frequency *everywhere*. The frequency will
be close but not exact, the many sets that supply power to the grid
are not connected by a hard physical link but by a relatively elastic
one of the long reactive grid distribution lines.

Not false at all! Every generator connected to the grid is phase-locked to
the grid and is thus bound to run at the same, grid, frequency.

I wonder what effect having lots of load that came on/off in response
to the (supposed) overall demand and supply ratio would have on grid
stability? With the time lag that it takes to bring ramp up supply
from coal/oil stations you couldn't really have stuff switching in
much less than 1/2hr IMHO and you wouldn't want all these things
doing a switch at the same time (a few minutes) relative to a
supposed dip/rise in grid frequency.

Sudden load changes cause a dip or rise in grid frequency. There are
tolerances on how much the frequency can vary and the controllers switch on,
or off, additional sources to keep the frequency within those limits.

SteveT