Too many variables to conduct a meaningful design of experiment. In a real world road situation, we feel and hear system stiffness in FD scrubbing noise, break drag, wheel flex. That system is a combination of wheels, tires, crank, pedal, seat post, seat post extension, stem, and bars. How do you come up with meaningful numbers that guarantee a meaningful trend over all combinations of variables?

IMO, stiffness is largely subjective, and may cause things that lessen someone's enjoyment of the experience. I had a BMC SL02, with full 105, and I got FD rub on every aggressive climb. I returned it, because what I considered flex was annoying. Now I primarily ride a Felt steel single speed, that is no where near as stiff as the BMC, but I get no noise no matter how aggressively I climb (because there is no FD ). Of course the BMC is stiffer, but the combination of variables made an intolerable condition for me.