On Fri, 12 Nov 2004 09:57:13 -0600, John Fields
wrote:

For a purely resistive load, since P = I²R, the sign of the current
going into the load gets squared, so that when it goes in negative it
comes out positive.

Seems like there's an opportunity for free energy kooks there.
Put positive and negative current in alternately, averaging zero,
but get all positive power. :-)