Originally Posted by
Fiery
This would be a pretty inaccurate calibration method since there is no way to be sure the ride has followed the one-mile course line exactly. The proper way to calibrate a cyclocomputer is to roll the bike down a marked straight line for exactly one wheel revolution (mark the tyre with chalk or something similar if necessary), and then measure the distance. For additional accuracy, repeat the process three or four times and take the average result.
Sorry, but Sheldon Brown disagrees with you. Sheldon Brown suggests using a measured course to
fine-tune (his words) your cyclocomputer setting.
http://sheldonbrown.com/cyclecomputer-calibration.html
I stick by my stated method.