The only problem I have with these "backyard" style of experiments is that they are not controlled. Personally, I'm pretty neutral on the whole debate. If you could manage to pull some legitimate stats that the thing works as claimed, good for you. If not, prove that it does. The biggest gripe I have with the "I'll fill my tank 5 times with and 5 times without." is the person behind the wheel. I can guarantee that my gas miliage will fluctaute from fill up to fill up just based on my driving habits for the last 2 weeks (I fill up twice a month). It depends on traffic, how I feel that day, the weather, the condition of my car, etc. Tons of variables to throw this experiment off the chart.
I don't think it will prove functionality for or against the product. A more controlled environment would be (afraid to say), the aforementioned dyno testing. A specific amount of gas, a consistant MPH, etc. Something that could be charted. My opinon anyway. But in these days with OMG HUGE!! mufflers and Type R stickers...hey, if you want to toss in an apple slicer into your engine air flow, go for it! ;)