The greatest variable in engine efficiency is how much of the available atmospheric pressure is being introduced into the combustion chamber.
Best efficiency will always be at moderate RPM with the lowest percentage of throttle position that allows each cylinder filling event to be as close to 100% of atmospheric pressure as possible.
The reason is:
An engine is basically a lever. It compresses a fuel air mixture, then ignites the mixture, which expands and applies pressure to the top of the piston.
Your best leverage is when you have the maximum possible pressure differential between compression and combustion (cylinder filling event).
The longest lever if you catch my drift.
Under higher loads you can get maximum atmospheric induction at relatively low throttle positions.
Using a vacuum guage, if you apply just enough throttle to get the lowest vacuum reading on your guage (while under some load) you have found the most efficient "sweet spot" of the engine. This will be true at relaitively low RPM, where the losses due to friction and reciprocation are at their lowest amounts.
A BSFC map clearly demonstrates this phonemon as well as illustrating that WOT is not the most efficient, when you consider throttle position. Throttle position can be as low as 30% up to 80% depending on the load applied.
This is why I always tried to stay in higher gears in my VX. You could easily check by adding more throttle, with no significant increase in power.
There may be specific examples where this does not apply perfectly in every situation, but as a general rule it's hard to beat.
In most cases, to achieve higher mileages, you have to make adjustments that are directly related to you specific driving environment. The best example in my case is the timing of the traffic lights on my normal route, which numbered over 50 in a 32 mile round trip.
If I accelerated too slowly I would catch almost every light changing yellow. This situation required me to accelerate more aggressively to avoid getting caught by the numerous lights.
I would still be interested if anyone could do a quick test with their scan guage to get an idea of what the unloaded fuel consumption of their engine would be at 2000 RPM compared to idle speed.
It would also be nice to see what horsepower was required to maintain both idle and 2000 RPM speeds using the same engine as the example.
My engine design eliminates almost 100% of reciprocation losses. It would be nice to be able to quantify the percentage of those losses compared to the cumulative losses of the same example engine, but the calculations include so many offsetting variables that precision may be impossible. This would allow basic efficiency calculations for a design that has never been built.