[Christ]

I said it's incorrect because even though the majority of your route was steady state, your FE reading would still incur hits from acceleration and deceleration, because you're calculating from tank to tank.

You, of all people, should know that a single tank is never enough to accurately decide whether something was an improvement or not, especially when tank to tank is your only testing method.

If you had some real time data, that you could read as you're driving, it would be a completely different scenario.

For posterity, I'm not going to argue the point. The information is on this site, as well as at least 3 others that I know of. If you're that interested in the data, you'll look it up.