I am not wrong. My "formula" is the definition of a Watt. If you end where you start from there will be no accumulation of potential energy. As for the drag factor, it's a loss that you can correct for. If I were going to do so I would determine my cross section and consider it a constant, and then measure the wind speed. Drag losses are a non-linear function of wind speed, but that shouldn't be a problem for a computer to calculate in real time.

Measuring the force applied to a pedal is a much more complex approach to this problem. The amount of force you apply will vary greatly with sprocket ratios. If you only want to measure force exerted that's fine, but if you want to measure watts it's going to be much more problematic.

Any approach will require some loss correction, to whatever degree of accuracy you desire. I think those corrections should rely on the simplest measurments (as in the most accurate, easily obtained and fewest in number) possible.

I don't know how negligible rolling friction losses (tires and roads, chains and sprockets, bearings and axels, etc.) would be. That would have to be the subject of more deliberation.

Joe