[aerohead]

From Donald L. Stivender, Engine Research Department, General Motors Research Laboratory, General Motors Corp, SAE Paper 780346
1) At a given constant throttle, essentially constant airflow is experienced.
2) At any given constant throttle position, the engine will exhibit a parametric BSFC ' fishhook', based upon F/A ratio ( oxygen balance ).
3) Engine torque = BMEP
4) Mechanical efficiency = BSFC
5) BSFC is inversely proportional to thermal efficiency
6) Optimum torque occurs @ ' Minimum Best Torque' - Spark Timing
7) BMEP is highly dependent upon A/F ratio
8) Minimum BSFC is defined by all throttle settings tangent to the throttle-stop fishhook
9) Fuel consumption- load characteristics generally apply at all engine rpm
10) Torque - A/F - efficiency, are all interrelated characteristics ( BMEP - BSFC - A/F )
11) Maximum torque occurs @ approx. A/F = 13:1 ( 'best power mixture' )
12) BMEP decreases with fuel flow at throttle - stop, torque per unit fuel increases to a maximum as fuel flow is reduced ( A/F = 18:1 )
--------------------------------------------------------------------------------------
13) Throttle position is the only link between engine operation and the CPU
14) Any change in road load is unrecognized by the CPU, as it's 'eyes' have been removed, via the Vernier Rheostat in the throttle position sensor fixed position.
15 The CPU cannot detect 'driver intent'.
--------------------------------------------------------------------------------------
16 Reduction in Road Load Horsepower forces the engine to shift it's BSFC map island, to that of a higher BSFC island, as friction effects now represent a larger fraction, increasing overall pumping losses.
17) A/F ratio experiences an excursion, taking BSFC with it, for overall lower mpg.
--------------------------------------------------------------------------------------
18) Aerodynamic drag reduction will precipitate the excursion, unless the original engine-map load is maintained ( gear-matching ).
19) Failure to maintain original engine load is the setup for the 'influence factors' reported by Gino Sovran, introducing 'hidden' losses within the vehicle's internal fuel management system. It's flying blind.
--------------------------------------------------------------------------------------
20) Any mpg attribution to an add-on aerodynamic appliance cannot be completely valid, by default.
Causality is indeterminant .
21) An 'observed' mpg improvement of say 0.6-mpg could actually be 1.0-mpg, only if the gear-matching is accomplished.
I've experienced this as a first-order-reality experience with 'Spirit.'
The highest mpgs ever returned were with 'rubber gearing', which reduced revolutions per mile, from 766, to 698. A $ 290 exercise.
A 'sixth-gear' would no doubt have helped even more. They don't make those.
--------------------------------------------------------------------------------------
I like the throttle-stop testing technique!
It covers so many of the published variables that, I'd be surprised if it didn't indicate aerodynamic trends.
As to the degree of precision we can expect, with respect to quantifying the actual aerodynamic performance of any given modification, Julian has already adequately addressed that issue.
The information I've presented is just to add some seasoning to a fairly tasty recipe.