A really good mileage booster that i have tried lately in my quest to get higher mileage is called Green Fuel Tabs. You can buy it online at GreenFuelTabs.com.

The nice part is that this is a tab, not a liquid, so it is easy to store in the glovebox, and small in size. No liquids to spill or clean up...my wife hates all the liquids I have tried and spilled. She is actually using the tabs now that I have shown her we are saving money with it.

Basically I got 10% improvement almost immediately, and the longer i use it the better my mileage gets...I'm up to 18% improvement now after a year of use...and I have a pretty heavy foot when I drive too. They also say they cut emmissions up to 50%, and I did see a significant drop on my last smog test, though I am not sure how to measure that. I use it for the performance and mileage savings. Check it out.

1st & most importantly - Acetone in your fuel system is a very, very bad idea. I like my fuel to stay inside the tank & lines. Acetone will degrade the seals and elastomeric lines in your fuel system. It will take longer at a low treat-rate, but it will take its toll. Acetone can also remove the easy components of a deposit and render the remainder intractable. If you feel compelled to add oxygenates beyond what is already there (if you live in a non-attainment area), go with an alcohol.

Next - there is a whole lot of smoke here:

'Not an additive, a treatment' - That is a semantic distinction that evades me at present. An 'additive' is something you add to something else; 'treatment' is the act of introducing an additive. That's why the amount of additive in a given formulation is referred to as the 'treat rate'.

'No chemicals' - Interesting, is this product not made of matter? Everything bigger than a neutron is a chemical.

As for the comparisons of before-and-after mileage - Would you like to buy a bridge? A human who just spent good coin for something that is supposed to improve fuel economy is going to want it to work. As such, that human is very likely to unconsciously modify their driving habits. The gains described here can easily be achieved by employing a few of the the less intrusive techniques from those known collectively as hypermiling.

Show me data from six cars, three with and three without, all run on the same dyno. Maybe we should go for 12 so we check both highway & city simulations. Then you'll have enough data to determine if there is any effect beyond normal experimental variation. (Fifteen of each would be required for a true statistical determination of the magnitude of an effect, but three will tell us if anything real is going on.)

Now all you have anecdotal data, also known as 'bedtime stories'.