close but no cigar..
started out encouragingly... yep a parabolic reflector would be good, then blam you completely fuck it all up with " if you apply pressure whilst polishing a flat surface you will induce a very slight curvature."
how much pressure, in what direction, for how long, on what thickness of substrate, youngs modulus of which is what?. cos you get any number of curves, only one of which will be parabolic.
(i'll give you a mulligan on the 'bond a bolt to the back of a mirror' - we'll assume cyano acrilates were avaialable to the greeks and they did have the capability of producing _perfectly_ flat sheets of glass or reflective metal with a _totaly_ uniform cross section.... ho hum)
but im afraid "The focal length of a parabolic mirror does not change its effective concentration" is pretty unforgivable. the POINT of a parabloic reflector is that it takes the energy impinging on the whole surface and delivers it to 1 point - the focus (same ammount of energy over a smaller area), passing beyond the focus the reflected beam diverges, therefore dissapating the energy, so that at a point that is as far from the focus on the opposite side from the reflector will see exactly the same concentration of reflected energy.
to reitterate: The reason this didnt work for mythbusters, or archimedes is that the whole thing is bollocks.
QED