True. Real harness engineering is when you actually measure outcome statistically. Although i would say 5 samples might be enough to conclude something