Duane -
Thanks for the note. I didn't post all of the columns from my spreadsheet, but at least 2 of the mile segments are off by more than 10 feet each! This is well outside the usual .08% tolerance.
I'm used to a variance on the order of 8-16 inches. I once measured a winding 2 mile course in a local park where the variance with the remeasure for the entire course was 1 count on a Jones counter (approx 0.05 rev).
I, too, am careful when stopping to ensure that a magnet is not close to the sensor. This is pretty easy to do with 2 magnets on the wheel - I just make sure one the magnet I see in front is near a horizontal position.
The RiegelSuperDuperCourseClicker (wow! wish I'd thought of that!) will give me an independent way to double-check the magnet/sensor combo, and it uses a mechanical technology. If the counter gets more than 40 counts per revolution, I should expect similar accuracy with it.
Your note about Cal #1 being off by a small amount seems reasonable, since I was using a 1000' course for all of this, but I haven't done the arithmetic yet to see how much variance to expect from it.