The faster cadence is technically less oxygen efficient, but since oxygen is not the limiting factor, that doesn't really matter. The reasons cyclists pedal at a fast cadence is because it uses the aerobic system more heavily than the anaerobic system, and the aerobic system is easy to keep "on" for a long time. You breathe harder, but your legs never hurt.

(I decided to start following this advice a few months ago. A 30 mile ride in 2.5 hours at 75-80rpm used to tire me out. My first 40 mile ride at a high cadence (95 rpm) took the same amount of time but I wasn't even close to being tired. My usual "training rides" of 20 miles take almost no effort unless I make a conscious effort to go anaerobic for a few minutes at a time.)

That's a lot more individual than the aggregate study results generally indicate. I'm a cranker, not a spinner, and I used to do 250 miles/week -- that included a "century Sunday" -- with little fatigue (no more than I was looking for, certainly). Most of the folks with me would be turning at about 150% of my rate at any given point. I tried the 90 cadence, and it just didn't work for me; I'd get pooped over a piddling 20-miler, and never really improved in endurance.

Yeah. It's important to keep in mind that the pros are pedaling their 100rpm in a 53/11 gear, whereas most of us non-pros are ... not.

When we race we adjust stride depending on the terrain- short strides on uphills level out the incline, and long strides on downhills let gravity take over and do the work for you.

YMMV (obviously), I find long strides going downhill far too damaging to my body. I practically fight gravity going downhill the way I do going uphill.