actually on second thoughts, perhaps this is mostly explained precisely by the extra motion involved with high cadence (running, cycling, probably even swimming). moving those parts of your body N times through space rather than (say) N*0.6 in order to cover the same distance has got to cost something … as you noted.
Correct…we can only measure pace or power, which is the output of everything the body does to put that mechanical energy out. What we cannot directly measure is the input energy. For example if I put my treadmill at zero miles per hour and do jumping jacks in place the measure power/pace is zero, yet I am getting tired. That’s the extreme case of moving your body around to do no work, yet it is tiring. Same as spinning you legs around with a snapped chain delivering zero watts to your powertap.
Years ago Carmichael gave a good example…if I have to lift 2000 lbs by squatting I can do it in 10 lifts of 200 lbs or 40 lifts of 50 lbs. In the case of the former, I end up moving my own 140 lb body weight 10x. In the case of the latter, I move my own 140 lbs body weight 40 times. So really the latter case is easier from a muscular force per stroke perspective, however, the total work is 30x140 lbs more (well it might not be the entire 140 lbs, but all the static upper body weight and some of the leg weight that you have to move upward from the squat position).
So yes, lower cadence is probably less overall work for you body, provided that the crank torque does not get so high that it takes the muscles quickly to failure. At IM race pace or RAAM race pace, the crank torque is not high, so the low cadence makes sense (see my posting above where I postulate that for most riders as power drops away from FTP, generally crank torque and RPM go down somewhat). Low RPM at FTP or higher (high crank torque) probably takes you close to “failure” after a lot less revs…that’s probably also why sprinters sprint at such high RPM versus having really massive gears (like 64 tooth pizza plates) and keeping the RPM down.
The human body is probably able to generate only so much force so fast. For example 100m runners, generally they say these guys take 4.7 strides per second (in the range of 47 strides for 100m)…that averages out to 141 RPM. I’ll have to go count the strides that Usein Bolt takes, but generally humans only produce so much ground force per stride at the very top end, when trying to maintain speed.
The ultimate in producing ground force is the long jump, but once they take that long jump stride, RPM goes down and overall speed is going down, even though the ground force was really high. So sprinting and jumping in running give us some good boundaries on what the human body wants to do from the perspective of force and RPM. Now strap us into pedals and things change somewhat, but I don’t think they change that much. We just adapt what we are naturally programmed to do but inside the constraints of the bike/pedal system.