I stated above that I always accelerate
much more rapidly than basically everyone else on the road (this is 6 stoplights per morning and 6 stoplight per afternoon 2-3 days a week at least when commuting by car.. and the hundreds of thousands of other stoplight accelerations ever since I began driving in 1995) and have
always returned better than EPA-listed mileage without using any deliberate hypermiling techniques.
All I do is get up to the speed limit fast, then stay there. I drew a crappy graph in mspaint to illustrate where I believe the difference lies:
Even if other drivers maintained the same maximum speed, they spend a much longer duration at acceleration-level fuel consumption, and a shorter duration at cruise level fuel consumption. I believe based on the feedback loop (with its necessary lag time) engines use to regulate combustion, a steady cruise rate will necessarily result in a more accurate and thus more efficient level of regulation. It's worth noting that this particular difference may not be relevant to electric vehicles because even if there is a feedback loop, it operates at the speed of light rather than at the speed of combustion gases.
Here is an example, not based on any actual measured data but just my
guess at why greater acceleration may not actually be a big culprit, even if the same cruise speed is attained - even when that means maximum cruise speed is maintained for a longer duration on the car that gets up to that speed first. Pay attention to the relationship between the green region and the yellow region as illustrated: