Maybe
You make some good points. Unless one goes to the same pump consistently to refill, there is additional variability in the calculations.
Is it way off or not depends on a number of factors. I would think today, in the age of evaporative emissions controls, which probably also apply to gas pumps, the pumps have to meet some standard of shutting off to avoid spillage and/or vapor release. So the pumps are probably more alike than different.
But, scientifically, the best, most repeatable results are when the same equipment and methods are used.
I was approaching the problem from the idea that we have no idea how each carmaker implements the algorithm that calculates the MPG. If they don't take into consideration idling, for example, they could produce an optimistic value.
My Mazda, which I realize is an anecdote and not data, is within 1% of the hand calculated value over 50k miles of driving. So I'm not sure I buy the "it's way off" camp's argument either.
And people are bad at math, so there is that...
Now let's argue about is empty really empty? I found out two years ago that even when my Mazda says something like 7 miles to empty, there is still close to two gallons on board, unless the published value for tank capacity is off by about 7L or 1.75 gallons
Originally Posted by itguy08
They are not off. The hand calculation methods are the ones that are way off. You have no way of knowing how much gas you put in the tank at each fill up.
For example. You drive 200 miles to 1/4 tank and fill it up to the first click and stop. You have no way of knowing how that pump was set for the click off (we've all had those pumps that click off near immediately and those that take a while). So while it may say "F" on the gauge and 15 gallons on the pump you have no way of knowing if it was just registering "F" or topped off. Next time you do the same and put in 16 gallons at another pump. Your hand calculations stated you got less MPG but that pump could have put you in a different spot on the "F" range (say topped off).
Many dash MPG calculators are a rolling type of the last 500-1000 miles. So they don't immediately change. I know when we tow it goes down like a rick but takes a few tankfulls to go back up.
Contrast that to the MPG meter which is set by the computer which knows exactly how much fuel is used for each injector pulse and can do the calculations much more precisely.
Take an average over, say 5000 miles and they should be pretty close. For my vehicles tracked with a phone app and the dash computer over the long haul (135k on one, 80k on the 2nd and 14k on the 3rd) they are all very close to the dash.