For these calculations I presume no wind, constant speed, constant
grade, racing crouch, smooth tires, and the power equations on page
157 of Whitt and Wilson, Bicycling Science, 2nd Ed., 1982,
Cambridge: MIT Press. Distance is measured as distance traveled (i.e.,
cyclometer distance) not horizontal distance, but for grades of less
than 15% or so the difference is pretty negligible.