There is a discussion going on of gray exposure with S-log. When using S-log the 38% rule does the trick imho.
I wonder what the sweet spot is when one records to the internal cards on cinegamma 1 or on the default profile (709ish). And is this the same with an 8-bit outboard recorder? (like the nanoflash)
On the EX series camera's I use around 50%, but content from the F3 always gives me the feeling that there is more to gain when I drop down a stop. Maybe this is due to the less noise in the shadows?
My guts makes me loose a stop in sunlight situations and use 'normal' exposure in dark scenes.
What is most important for me are the faces.

The text book would say 18% grey card at 50% with REC-709 or cinegammas. But if your going to grade or you just want a darker look you can comfortably afford to underexpose by a stop.

On an F3 set to REC709 settings I am usually setting my faces at 40 - 45 IRE to keep them from looking too video "bright". The camera is so clean that even if this underexposes the background somewhat, it still records without noise for the most part depending on the contrast ratio between FG and BG.