[cyclopathic]

It is funny how these VG threads degrade after several posts, dog mask, heh.

Back to aero posts: if we take for the basis numbers posted in Mitsubishi article (1.4% Cd reduction and 4.5% in Cl) detecting 1.4% outside of controlled experiment would be difficult. For example on my commute route MPG may vary within 10mpg depending on temperature, lights and traffic situation. 0.7-0.8mpg gain will be lost in the background noise.

I disagree that articles like in OP should be discarded completely. While they cannot claim any numerical accuracy, they still give useful comparison. Even if you can't say by what % A is bigger than B, you can still say that A is bigger than B.

There are numerous credible studies on VG airfoil application which prove that VGs can speed up boundary layer and delay flow separation. With cars having more complex shapes things get more complicated. Finding an optimal location for VGs is fool's errand or it could be.

Cars also need to adhere to aesthetics and safety regulations and should be easily incorporated into manufacturing process. So far the most common VG application is to the rear in attempt to reduce wake trail. I am not convinced they will not be effective in the front but may impose pedestrian crash hazard when applied to bumper, hood or roof leading edge.

But this does not mean that VGs are magical item which only works when placed in right place. There could be still benefits even if they are applied to less than optimal position.

Also last but not least assumptions on point of flow separation could be wrong. Even in relatively simple case (hatchback) you can't just assume it will be roof trailing edge. It could be a gap between the roof and hatch door.