A Florida reader questioned the gas mileage that I reported for Honda's minivan, the Odyssey: "Don't you feel the least bit silly," he wrote, "telling people that this car gets 28 mpg highway? If it gets 23 at 70 mph I will eat my hat." I wrote back that the vehicle gave me 23 mpg in town, and on the interstate the Odyssey averaged 28 mog -- or so its trip computer informed me. I was going to finish with something like, "and who doesn't trust Honda?" but then we all trusted VW, and look where that got us.
I have absolutely no reason not to trust Honda, a company that, in addition to cars and bikes, produces brilliantly engineered power tools ranging from snowblowers to jet planes and has been a leader in gas-squeezing and exhaust-scrubbing, but let us consider reported vs. claimed vs. actual gas mileage.
The reader added, "Try cross referencing against the info you get from filling up at a gas station and I am sure you would find the onboard computer to be very optimistic." But for me, trying to nail down fuel efficiency to tenths of gallons per mile becomes an exercise in frustration. For starters, do I trust the gas pumps? (They're covered with state inspection stickers, so they ought to be accurate -- but how accurate is that? See below.) Then I need to know how far I've driven -- but if we're questioning the car's trip computer, why would we believe its odometer? Instead, I should map out highway and city routes to the nearest fraction of a mile. Since we're distrusting the car, I guess this would have to be done by GPS, which now is generally accurate to about 15 meters, or 50 feet.
I live at sea level, so altitude wouldn't be an issue -- except to readers who live in the Rockies. What about hills? What goes up usually comes down again, so their effect on fuel consumption cancels out. Tire inflation is easy to manage, and these are new cars, so their motors should be in good tune and their alternators sending the correct voltage to the instruments. Finally, the driving itself should be normal, consistent and at the speed limit -- oh, but then there's traffic. And headwinds, tailwinds, heat and cold, rain and snow, which also play a role in fuel burn . . .
This where I throw up my hands and say, Screw it -- the real world is too complicated. In the old Mobilgas Economy Runs, which ended in 1968, cars had calibrated mileage wheels attached and their measured fuel was stored in special flasks in the trunk. Without all that precise metering, I go back to the car's onboard computer.
Still, shouldn't tracking fuel burned against miles driven as shown by the pump and odometer at least help check the car's computer?
The federal standard for gas pumps is a maximum allowable error of 0.3 percent. This means that, by law, an indicated 10-gallon fill-up could be as much as 10.028 gallons or as little as 9.984 gallons. The government regulates car odometers too, to plus or minus 4 percent of actual mileage: After a 100-mile drive, our odometers can legally show as many as 104 or as few as 96 miles. So depending on how these tolerances stack up -- 9.984 gallons for 104 miles or 10.028 gallons for 96 miles -- we could see a variation of as much as 0.9 percent in mpg between two law-abiding models of the same car driven the same way over the same roads in the same weather with the same tires, etc., etc. In a 14-gallon tank, that's just two cups of gasoline. Not bad, on paper.
However, along with all the above variables, I've found estimates that as many as one in five gas pumps are out of whack. Also, how fast or slow we pump the gas can vary how much we actually get, compared to what the meter says; and, compared to the odometer, our actual distances driven often changes when we replace tires because of differences in tire diameters. Again, aagh!
By averaging pump and odometer figures from a dozen gas-tank fill-ups, we are probably able to come within 4 or 5 percent of actual fuel efficiency -- which means that a calculated 30 MPG might really be only 28.5 or as much as 31.5. I'll wager that this is no better and probably worse than the accuracy we get from a new car's trip computer -- but I will continue to write "indicated" or "according to the computer" when I post my own mpg for a particular car. Just so you know.
The feds warn us that "actual results will vary for many reasons, including driving conditions, how you drive and the condition of your vehicle," but once in a while a car gets into hot water for consistently delivering way worse mileage than its EPA/DOT Fuel Economy Estimate. This happens because drivers pay attention to their cars' trip computers and/or odometers. So I always wonder why -- if the intent was truly evil -- the carmaker didn't also jigger its cars' electronics to show numbers that back up its bogus claims. If you're going to cheat, do it right, right? Sooner or later, though, cheaters get caught. As we've just seen in Germany.
Silvio Calabi reviews the latest from Detroit, Munich, Yokohama, Gothenburg, Crewe, Seoul and wherever else interesting cars are born. Silvio is a member of the International Motor Press Association whose automotive reviews date back to the Reagan administration. He is the former publisher of Speedway Illustrated magazine and an author. Contact him at calabi.silvio@gmail.com.
