I don't know if you are the "Chung" responsible for the Chung approach to this type of analysis or not ( Blather 'bout Bikes: Aero Field Testing using the "Chung Method" - How sensitive can it be? ). But I started my stuff before I encountered this.

But fundamentally it is no different. You know the power in (by the second) and velocity (by the second). So assuming that you know the net gradient of the route, Energy In (known) = Energy Out (known) - Losses (can be calculated given a Cda and Crr-drag and rolling resistance coefficients). The Chung method assumes a closed loop (so no net energy effect due to gravity, regardless of how hilly the terrain). I tend to use routes that are not closed, but subtract 'pairs of runs' which cancels out whatever net gain/loss of altitude exists. But the same thing. Fundamentally a far less elegant implementation of what is available in Golden Cheetah.

I also have a bike simulator where you input your route (slope is the key parameter), Crr/Cda, and some relationship that outputs power to see the net effect of various changes. This implementation rivals my previous one in terms in not being elegant :-)

dave

ps. One of the disadvantages of using pairs of runs (subtracted) is that you really need to be using pairs of runs with reasonably different power profiles. But I admit that I ultimately didn't do much with this. Have done more with the bike simulator.