Originally Posted by RChung

Yeah. I think Ballard does pretty good analysis, though I was surprised that, in that video, he said that he was surprised about the result. I thought that was well-understood even if not always well-known. We'd done experimental tests of wheels with different rotational inertia on the track (for team pursuit, where accelerations/decelerations happen all the time). It turns out that humans on bikes don't accelerate very quickly (in an absolute sense) even during a track pursuit -- on the road, or up a hill, or even in a crit, our accelerations are even lower. So the mass of the wheel matters but the contribution to power demanded is *almost* *entirely* described by the additional mass term, and not the moment of inertia term. That's why in my estimation equation for drag, I usually leave it out. There's a new paper in Sports Engineering that models the power demands of track cycling. Their model looked at instantaneous power demand around a track, when your wheels are accelerating as they do in the turns or decelerating as you exit the turn, but I look at net power demand (where you have to look at both the acceleration and deceleration, net) across a lap. My recollection is that, while they do look at a rotational inertia term, its contribution to total power demand when you integrate across a lap is essentially zero.

Light wheels and good tires feel nice, though.