The author of that piece tried very hard to make it look very scientifical, but then ruined it with this snippet: "Based on the model developed in Appendix A, and data I have collected while riding (velocity and elevation profiles),
"...data I have collected while riding" means that his n=1 -- and, since he was not in a lab, he could not possibly have held other variables constant.
That ‘data collected while riding’, refers to the speeds and elevation profiles. That is perfectly fine. That is not an experiment, where sample size is indeed, that is just creating a profile of a ride that you are modeling. If you want to calculate it for a different terrain or riding speeds, you can enter those numbers in the model, and calculate that.
I don’t think he did any testing himself.