Some things here leap out at me.

Using hour record attempts as a template isn't a great idea. The style of riding between an outdoor bike split and a velodrome are completely different. There have been a comparably low sample rate as well. 20 odd riders over a 100 years isn't ideal.

Is comparing pro's to age groupers wise ? The top pro's are likely to be physiologically suited (its unlikely they would have been pro's otherwise in the first place) whereas MOP riders will have many 'faults' that may require differences in set up or riding style. Unlike a top drawer pro, if a MOP rider has a comparably low Vo2 max, would it not make more sense to ge them to ride a lower cadence whereby the Vo2 max strain is reduced ? (excuse my wording on that one)

Cadence is only one part of the human/machine interface. Trying to isolate an 'ideal' whilst ignoring power applied/ muscle fibre physiology/ crank length & bike fit angles seems like we're only looking at the topping, not the whole pie. A similiarity in one characteristic doesn't make it right or statistically viable.