RChung

Right, I think cadence is a dependent (aka endogenous), not an independent (exogenous), variable. I think riders decide the speed they need to keep up in the race and that implies a certain amount of power. They can achieve that power either by changing their cadence or their crank torque or both -- but in the data I've seen they *mostly* modulate their power by varying their crank torque.

Here's a plot from a different rider on a different hill. The hill was about 15 minutes long, and I snipped away the rest of the ride to focus only on that hill. The top two panels show the relationship between power and cadence, and power and crank torque. The two bottom panels show the relationship between cadence and torque. As it happens I can see his choice of gearing at each second of the climb, and the lowest ring-cog combo he used was a 39/23 although he had a 39/25. I think you can see that the relationship between crank torque and power is much stronger than the relationship between cadence and power -- that means this rider is *mostly* using crank torque to modulate his power.

The bottom right panel shows you how the combinations of cadence and crank torque he used were related to his power.