I'm with Jeff here. Logic dictates that a longer test-run might minimize the 3 run-time MPG variations, AND be more representative of the MPG being achieved.
I just had a 65 MPG segment over 30 miles, but have never even had a 50 MPG tank (yet). So I'm not concluding that this one will hit 50 (let alone 60).
I also suspect that mostly level road may hav been favorable to the tailwind portion of your test.
What you are attempting to measure IS hard to definitively do, without being in a clsoed enviromnment. I think it's a great experiment though, and applaud your enthusiasm
