Hi John,

I think we have a likely culprit as to why the satellite magnitudes
might differ. Kevin Fetter kindly mentioned a page describing how the
absolute magnitudes of satellites are computed:

http://www.satobs.org/brite.html

which describes a difference in the magnitudes from the Molczan and
Quicksat files (see text below, copied from above URL). It looks as
if I should just knock the Quicksat magnitudes down (fainter) by 1.5.

At least in theory, I could probably compute magnitudes based on
radar cross-section. The benefit of this would be that RCS data is
available for almost all objects, whereas a lot of objects lack either
Molczan or Quicksat dat. The drawback would be that RCS correlates
only loosely to physical size, which in turn correlates only loosely
to magnitude (that is, some satellites may be covered in black paint
and so forth.)

-- Bill

"...The "Molczan" magnitude values are based on 50% illumination; the
"Quicksat" values are based on 100% illumination. The difference
mathematically is 0.8 magnitude. The "Molczan" magnitudes are 0.8
magnitudes fainter due to this assumption.

The "Molczan" values are based on "mean" magnitudes, while the
"Quicksat" values are based on "maximum" magnitudes. Typically (for a
cylinder) the "mean" magnitude is about 0.7 magnitudes fainter than the
"maximum" magnitude. On average, the cylinder is tilted about 45 degrees
relative to the observer compared to its maximum possible attitude.

The sum of the two differences in magnitude is about 1.5 magnitudes.
So a typical Cosmos rocket would appear in the Quicksat magnitude
file as intrinsic 4.0 and in the Molczan file as intrinsic 5.5 , that
is, its value of brightness in magnitude would be dimmer compared to
the Quicksat value."