Profile averages seem off

[NBtarheel33]

On the profile page, I believe there may be an error in how some of the averages are calculated.

I have 97,726 total miles tracked and 324 total fuel-ups. This would seem to imply an average of 97,726/324 = 301.6 miles per fuel-up, yet the profile page is quoting an average of 303.5 miles per fuel-up. Since 97,726/322 is roughly 303.5, it looks as though Fuelly may be using a fuel-up count that is short by two in calculating the average fuel-up distance.

Also, the average price per gallon figure differs from what I have in my spreadsheet. My calculation is based on total spent divided by total number of gallons purchased (i.e. $8,052.23/3,067.492 = $2.6250/gal), while Fuelly is reporting an average of $2.64/gal. Perhaps this is because Fuelly simply averages the prices of each fuel-up, rather than keeping a true running tally as I have in my spreadsheet?

Note that the average *total* price per fuel-up *is* correct. I have $8,052.23/324 = $24.85, which is what my profile page is reporting as well.

Anyway, I was just curious about what was happening here. As always, thanks for all your work on the site!

[pb]

You're using odometer tracking, so keep in mind that Fuelly can't use your first fuel-up for its calculations. It needs a baseline reading to start from and from that moment on it starts calculating the distance traveled with each fuel-up. Using that first fuel-up without a distance would throw the average off. That brings us down to 323 fuel-ups we can use.

You have two fuel-ups on Feb 12, 2010 that have the exact same odometer reading. I'm guessing that was a mistake, and once you remove one of the duplicates that will bring the number of fuel-ups to the right number.

And yes, for average price Fuelly does a simple average of all prices instead of total spent / total gallons. We calculate average MPG using totals, but not prices. We could probably move to that for prices too.

[NBtarheel33]

Actually, the two fuel-ups with the same odometer reading are legitimate distinct fuel-ups. I actually had run out of gas, so I had to put in just enough to get going again and limp under a mile to the nearest gas station to fill up the rest of the way. Turns out that the odometer reading was the same at the point where I had run out of gas and at the gas station. Is there a way around *that* kind of situation?

Also, with respect to overall average MPG, Fuelly shows 32.1 and I have 31.97. I'm guessing that this is due to the difference in averaging methods, or due to one or two fuel-ups being thrown out.

[pb]

Yeah, the average miles/fuel-up is going to be off from your method of calculating when you have two fuel-ups with the exact same odometer reading. Fuelly can't calculate a distance if there's no difference. And if there's no distance, there's no way to include it in the distance-related statistics.

[NBtarheel33]

Why not calculate average miles per fuel-up as total miles tracked divided by n-1, where n is the number of reported fuel-ups? This would still throw out the first fuel-up, while simply incorporating "no mileage" fuel-ups as 0 miles (which is really what they are). Obviously, this will make the average miles per fuel-up lower, but the same sort of thing would happen even to someone who fueled up every 20 or 50 miles, right?

[pb]

People sometimes switch tracking methods between tripometer and odometer, so there's no guarantee that n-1 would always be accurate. Zero miles/fuel-up feels like an edge case to me. We want people to enter fuel-ups with a distance and anything else is going to confusing and hard to explain.