I might have just solved the question about why the FCD always seems to think we're doing anywhere between 0.5 and 2.0mpg better than hand calculation. Hand-calcs, by their nature, are always one tank *behind* what actually happened, because we haven't taken into account the *factory fill*. When I added in another 12.5 gallons "used" to account for this, my split went from +0.73 in favor of the FCD to +0.01. Folks who have been tracking every tank using both methods, see if you get the same

 

I am not entirely sure I understand this post, my display consistantly reads the same or lower than hand calc, usually reads lower (display 42, hand 44) (follow the link in my tag, I have documented each tank.) But what I am seeing you say is your display reads higer than han calc (45 display, 42 by hand?).

 

I think what needs to be done is to compare the FCD of the current tank with the hand-calc of the previous, assuming, of course, that you're consistent in your fill methodology. There will still be some deviation across any single tank, but over the lifetime of the car, well, I just whacked off 0.72mpg of "error" by remembering that the factory had to fill my tank, too.

 

But the factory doesn't fill your tank. They just put in a few gallons.
GL

 

I don't really understand what you're saying. If fill my tank full, drive 600 miles, and then it takes me 11 gallons to fill it back up to full, I drove 600 miles on 11 gallons.

When I bought the car, they filled it up (maybe not COMPLETELY full, but full enough to hit all bars). I drove the car for x-hundred miles on that tank and then re-filled it back up and calculated based on the amount of gas I put in. If I added another 12.5 gallons, the calculation would be off since I calculated that after my first fill up.

I think it's just a strange coincidence that your numbers come out to 0.01....

 

Gordon: Dunno about your dealership, but one of the advertisements at the time I bought mine was "We send you out the door with a full tank."

kristian: You say, "If fill my tank full, drive 600 miles, and then it takes me 11 gallons to fill it back up to full, I drove 600 miles on 11 gallons." That's right, sort of. The problem is that (at least in my case), the factory/dealership fill wasn't being included in the calculations. That is to say, the calculation was based on "I drove 600 miles and then filled with 11 gallons." Instead, it should be, "I filled with 11 gallons, and got 600 miles before I stopped to refill." By changing the order of the sentence, I've also changed where the addition/division takes place. Admittedly, it is just a theory on my part, but that's why I'm asking for other folks who have been keeping track of both FCD and fill the way I have. If they also go from "very optimistic" to "practically spot on", we're on to something. If not, well, strange coincidences are what makes this world so much fun

 

Would this method take care of the "factory fill"? Each fill I compare hand-calc vs gauge. My hand-calc is consistently 4% higher than the gauge.
To address the statement that one is always one tank behind, I have computed my lifetime hand-calc mph (total miles on the car / sum of all gas I have put in the car). I have compared this to the average of the gauge reading for all tanks. My lifetime hand-calc is 51.3 and the average of my all my per tank gauge readings is 49.5. These two numbers correlate well with the statement above that my hand-calc (per tank) is consistently 4% higher than the gauge reading (per tank). The car presently has over 22,000 miles.
To me it looks like my hand-calc is going to be about 4% higher than the FCD no matter how many miles I put on the car. Does anyone else have data showing hand-calc vs FCD differences over a large number of miles?

 

I'll post a link to my Excel sheet when I get home so others can play. I'm currently just shy of 47k miles.

Edit for linkage

 

This seems like a silly discussion. Just compare the FCD to your hand calc for that same tank!

Ignore the first tank (or top it off and reset your FCD).

 

Two problems with this theory:
1. Airplanes seemed like a silly discussion 105 years ago, until a couple of overachieving (or maybe just plain bored) bikers proved otherwise.
2. The variation between tanks using your method, with just my data and not including partial fills, can be anywhere between +4 and -3. Useful trends only emerge after the data set grows large enough to wash out the noise.