Not claiming to be a top rider or athlete, or claiming to be an expert - just floating some figures to contribute to the ‘constant foot speed’ discussion.
Tonight I took the 28 out for a bit of a ‘burn’ and took readings off my computer after 1/2 hour of riding briskly, mostly on tarmac, but with some simple cross country sections. I have ridden similar routes on my 24 and Coker and recorded the half hour figures. then I rode for a further half hour or so ‘as the fancy took me’.
24 inch Nimbus. 102mm cranks (4.02 inches) 12 inch wheel radius. Ratio: 33.46%. Top speed 14 mph. Best average speed over half an hour was over 9mph but under 10mph.
28 inch Nimbus. 110 mm cranks (4.33 inches) 14 inch wheel radius. Ratio: 30.92%. Top speed 15mph.
Coker 125mm cranks (4.92 inches) 18 inch wheel radius. Ratio: 27.33%. Top speed 14 mph. Will average 10 mph over half an hour.
Nimbus 28 did 5.11 miles in 30:20 = 10.1 mph average.
So, the constant crank speed hypothesis, and the ‘gear ratio is the prime determinator of top speed’ hypothesis are not strongly supported (middle uni of the three has the highest top speed).
These figures are not scientific as I did not follow an identical route in identical weather etc. but they are a guide. Interestingly, the top speed on the 28 was recorded during the mixed riding phase. On the 1/2 hour blast, it recorded only 12mph.