(I see that you properly "take things with a grain of salt", and also
question simple answers for questions that are not so simple as
they may seem...;-)

With a steady hand and multiple exposures of the same thing the
1/FL speed rule can work (or even with slower speeds...), but one
complication you catch, below...  Also, the *rated* improvement
offered by stabilization probably usually has some ad-hype "slosh"
built into it and is overly optimistic...

It is non-linear. Try shooting an 8mm fisheye on a FF body - SCADS
of hand-holdability are there at rediculously slow shutter speeds, well
below 1/8th second. Even the 24mm is easy to hand hold at 1/15th - but
a 135 is a bit challenging at 1/125th and a 300mm is more than a
challenge at 1/250th (it is nearly impossible). But while most good teles
are sharp wide open, most wides are not and can benefit from much
smaller stops (especially in the corners), so the practicality of things can
even out somewhat...

While this should be true, it does seem that smaller sensors not only
produce greater apparent DOF, but fewer problems with hand-holding.
Not sure of this, though...

Yes, unless there are disadvantages...

Prolly not.......! ;-)

--
David Ruether
DRuether@twcny.rr.com
rpn1@cornell.edu
http://www.ferrario.com/ruether

Nikon claim *** 4 *** stops with the VRII on their new 70-300 4,5-5,6 tele.
That means, what, -4 stops from 1/450 (1.5 crop) = 1/30th handheld
Somehow I don't think so.

What manufacturer states anything like  your assumed rule of thumb?  I
believe all the claims to be relative, not absolute.

Kind of an odd coincidence, eh? Yet, surprisingly, it sort of works, at least
up to about 200mm. Beyond that, it breaks down rapidly. At least, that is how
it works for me. I find it very difficult to get a sharp handheld shot with a
400mm lens at *any* shutter speed.

You adjust the rule of thumb to your own preferences, really. No one can
decide for you how sharp is sharp enough. Any handheld shot is likely to be
softer than one with a tripod. Considered another way, many people complain
about the way depth of field markings have disappeared from modern lenses.
But those were just rule of thumb guidelines, too. There is an exact focus
distance. Anything on either side of that will be softer. Some people want to
be told how much softness is acceptable to them, I guess. Others figure it
out on their own. Personally, I just use depth of field preview when I am
interested.

So, experience teaches you to modify your rule of thumb.

Can't really speak to that. It seems obvious to me.

I think the idea is that you will get the same level of sharpness three stops
slower with IS than without it. As a practical matter, I think this is
conservative. It depends a little bit on how much your shakiness increases as
time passes.

Competitive marksmen are very familiar with this concept. Your aiming point
is never still; it moves around in a roughly circular area (actually a little
wider than it is tall) in a kind of figure 8 pattern centered on the target.
For a short period of time this circular area remains a constant size. But as
you continue to aim, the circular area will begin to grow. You can reduce its
size briefly by holding your breath, but within moments after you begin to
hold your breath tiny spasms throughout your body will again increase the
circle of error to an extremely large size. This is why you do not hold your
breath until the moment you fire -- you need to be already squeezing the
trigger. Then you breath again immediately.

Image stabilization keeps the circle of error smaller and of a constant size
-- up to a limit. Eventually your body will begin to have small muscle spasms
as a reaction to trying to hold still. These spasms will occur sooner and be
stronger if you hold your breath, but breathing introduces even greater
errors. But think about this -- as long as the circle of error remains
constant and the IS is adjusting for that much error, every shot will show
pretty much the same degree of sharpness, no matter how long the exposure is.
The only limiting factor is how long you can keep your aiming point within
the tolerances of the IS. Photography demands, for the purpose of sharpness,
movement that can be measured in angstroms. You do not have the same margin
of error that you would for a comparatively gigantic bullet. (Tell that to
your hunter friends -- good photographers have to be better shots than they
are. :-) )

The way this works out in real life is that it seems to me that I can hold my
circle of error constant for about 1/8 second. A handheld shot with a VR
400mm lens for me is no worse at 1/8 second than 1/1000 second. Trouble is,
either one is pretty atrocious. Good thing I didn't become a brain surgeon.

than others, depending largely on how tired I am.

So that's my take on it. Manufacturers' claims are not really comparable to
real stops of improvement. It works differently, and it seems to work a
little better than what the manufacturers say it does.