[Vanner]

I feel very argumentative at this point, but I finally flipped through the articles, and the author mentions fuel economy in only one sentence: a "gut feel" of "fractional" improvement that he himself admitted was "not very scientific." But that's fine, that's what I would expect from an article about improving performance from a magazine called autospeed.

As for your results (and the magazine's for that matter), I'm not saying they're inaccurate, but unfortunately there are dozens of variables affecting fuel economy. Outside temperature, the particular road used, speed traveled, weather, traffic, the driver, accessories, cetane/blend of fuel, other modifications done on the car at the same time as the CAI, and adaptations performed by the ECU could all have major, major effects on fuel economy between tests. With all other variables accounted for, I just simply do not understand how better economy is possible from a theoretical standpoint outside of the exceptions I presented above.