From what I understand the reason for the value of 5 has to do with the fact that fact that the BTZS recommends and uses a incident meter for it's application. The hemispherical dome on the incident meter averages values to that value of 5.
With an incident meter, one measures the highlight and shadow values as they present themselves to the light that is striking them. Not as the light from those areas are being reflected back toward the camera.
Additionally the conversion from a reflective reading (spot meter) to incident meter is not a truly linear conversion. I would guess that a linear conversion would get one somewhere into the "ballpark". The BTZS has a conversion formula for reflective to incident readings.