As I understand, the required flash power for correct exposure is determined by aperture, ISO and subject distance. It shouldn't matter whether a human subject is dressed in black or white -- the required flash power should be the same.
With this in mind, why is it necessary for ETTL to fire a pre-flash to determine the correct flash exposure? Further, why is it necessary to apply flash exposure compensation for subjects that aren't midtones?
The camera doesn't know the distance of all subjects - only a rough estimate of the distance to which the lens is focused. The camera also does not know the exact flash head angle, it may be the case that you are bouncing the flash of the ceiling, or there may be other surfaces in the scene that reflect light back onto the subject.
Flash exposure compensation is required exactly because the camera doesn't know the colour of the objects in the scene, and therefore how much light to expect back. It can only make the standard metering assumption of 18% reflectivity.
Firing off a preflash and measuring light returning through the lens is still the best way to judge the required flash output for a correct(ish) exposure.
Even if you knew the distance (some cameras will transmit focus distance info to the flash system), you still don't know how much reflected light is being bounced back on the subject. The flash exposure needed would be greater outside than in a small room with white walls.
