Some numbers awhile back bugged me a bit... the quote from the OP is ~ "15-20% of the input energy at an electrical power plant was used at the final destination".
Ok. But, then:
"figure 33% of the fuel energy can end up as mechanical energy at the flywheel for a 4cyl Honda"
Hmm. Both statements are apparently true. But I think the comparison is unfair. The electrical efficiency assumes all losses incurred (conversion, transmission, re-conversion to useful work), essentially an end-to-end efficiency. The automotive engine gets 33% (arguable, but maybe possible with today's turbocharged, multi-valve motors) only at its best operating point, which is essentially near-redline with near-wide-open-throttle. Since that is not how most of those Hondas are operated, the efficiency of the engine is significantly lower in actual driving use due to operation at part-load, part-throttle "cruise" conditions (maybe more like 20-25%? But I'd bet a bit less than that even). Also, to be fair, the auto example should add in conversion losses thru the drivetrain, reducing the figure to something closer to 15-20% for an overall efficiency. These are just WAG numbers, obviously, since actual drive cycles could (and do) cause a lot of variation in the average efficiency of the vehicle.
It's always bothered me that fuel-burning locomotives actually run the engine to generate electricity, which is then applied to the electric traction motors to move the train... But it's obvious when you think about it, a DC motor can provide ample torque at zero speed, while a diesel engine would have to have a relatively inefficient torque converter with gobs of slip to allow it to move the train from a dead stop. But why lug the weight and fuel around for the diesel? Was told long ago that weight for a locomotive is beneficial, since it provides traction... but that arguement does not hold for passenger rail systems where all cars have drive wheels, or are so light that high traction force from the loco is not required. I think the real answer is that diesel used to be really, really cheap. So, no point in installing expensive overhead power lines for cross-country travel.
If electrical power is locally generated, through a distributed network of small powerplants (of whatever ilk), are transmission losses across the grid consequently reduced? Does that argue in favor of wind/solar/small hydro/farm methane etc. etc. generation?
Hmmm.