If the density of stars is constant (and observations show it roughly is on large scales), the number of stars is proportional to the volume. The volume of space within a distance r is proportional to r3 (4/3 pi r3).
But you're not concerned about volume, you're concerned about distance. This would equate to the surface area of the sphere, with r measured from the centre of the shell.
This shell area is proportional to r
2. (This should make sense, since the inverse square propagation of light is just the inverse of the area the light has to move through)
It still leaves us with a constant though.
But, as Parsifal pointed out, stars can't multiply up their power, which seems to be an underlying assumption.