The question remains: would you ever be able to detect any meaningful differences between, say, the Perez All-Weather (or an overcast/clear blend) and the 15 skies of the CIE GS formulation? This paper suggests not: Sky model blends for predicting internal illuminance: a comparison founded on the BRE-IDMP dataset. Journal of Building Performance Simulation, 1(3):163–173, 2008.
Recall also that the daylight illumination received anywhere in a space is always some mix of direct and inter-reflected light from the sky and (when present) the sun, and never that just from a patch of sky (unless the space is pitch black). Accordingly, the effect of subtle (and not so subtle) divergences between sky luminance patterns on the internal daylighting conditions gets ironed out, often such that any differences remaining are likely to be only of (ahem) academic interest.
The CIE 171:2006 tests are crude, and there were some errors in the original.