That's one heck of a pull . . .

[MattKing]

Just to throw it out, here is a related observation. This only applies to a print made once one knows how to make a good print that matches the negative to the paper grade.

Under exposed negative: When shadows printed to go to black and highlights printed to go to white the middle is too dark (scooped out by the toe)
Over exposed negative: When shadows printed to go to black and the highlights printed to go to white the middle is too bright (pushed up by the shoulder)

That form of over exposure is extreme (5 or so stops) and gives an odd look to the print. In my experience, this odd tonal scale, is responsible for the drop off in print quality way over to the right on the "Contact Print" line in the graph shown in post #2 above.

FYI: Another way to look at it:
Under exposed negative: When the middle and highlights are printed correctly, the shadows are gray. This looks really bad. This is how novice prints frequently come back from the lab.
Severe over exposed negative: When the middle and shadows are printed correctly, the highlights are gray. This looks really bad also.

This is a really good way to describe things. IMHO, in these exposure discussions not nearly enough attention is paid to the middle.

 

[ic-racer]

That graph is just a model, but the empiric data depend on a lot of variables. For example, my Nikon with state-of-the-art Matrix Metering has a narrow probability curve, such that the safety factor for a 2% to 5% under-exposure rate is around (0.5 Stops). That really is 'space-age' technology; I shoot almost everything on "Auto" with those cameras.

That is in strong contrast to my results from large format photography. In that case the bell curve is wider. Likely due to errors in spot readings (flare, accuracy of the spot), low zone determination (is that 2 or 3?), shutter speed variability (20 year old Copals, etc), aperture setting (no detents on any shutter), bellows factors, film age speed depression, etc. In that case my safety factor is around (1.3 Stops).

So, even though I have all the testing equipment, I'm NOT the guy that tests and records exact shutter speeds on all 20 shutters, does in-house ISO determination on every box of sheet film, carries a ruler to set bellows factor for every shot less than infinity focus, uses a ground-glass exposure meter calibrated by the factory on a bi-annual basis, etc.

Am I sloppy? I don't think so, I think I'm smart enough to know what I can get away with.

 

[Berkeley Mike]

Cool graphs. The images in the article the tones-to-whites seem to be blocked-up. As far as the article either the images suffer from publication issues (mags/books copied or digital artifacts.) That's what happens when you over expose. I see people liking the result. I just don't get it.