Thanks much for the report and the link to the Zero test. I have a different interpretation of the test. The test posted below looks rigorous enough but the shortness of it means the EPA had to extrapolate the results by more than 15x to arrive at 114. That creates a large margin of error that may explain the differences between the EPA estimate and riders' actual experience.
The test subjects the bike to 18 starts and stops over a period of 23 minutes over a distance of 7.45 miles.
1. Hard acceleration to 20MPH then acceleration to 30MPH over 90 seconds, then rapid deceleration to zero.
2. Hard acceleration to 37MPH, then again to 56MPH and coast over 60 seconds, then rapid deceleration to zero.
3. Hard acceleration to 35MPH, then rapid deceleration to zero.
4. Hard acceleration to 30MPH, then rapid deceleration to zero.
5. Hard acceleration to 35MPH, cruise for 30 seconds, then rapid deceleration to zero.
6. Hard acceleration to 25MPH, then rapid deceleration to zero.
And so on. So, yes, the average speed was 20 but the way the bike achieved that was via many power-intensive rapid starts, accelerations, and stops. Maybe the test used 5% to 10% of the total charge. If they started with the assumption that the 7.45 mile test consumed 10% of total charge then the total range estimate would have been 75 miles or if they started with 5% then 20 x 7.45 = 149 miles. The 114 mile official estimate implies that the EPA extrapolated total range per the test above consuming ~7% of total charge for 7.45 miles times 15.3 = 114 miles at 100% discharge. The methodology produced a wide range of error.