Final Inspection

When It Comes to Mileage, Cars Tell the Truth


A vehicle’s trip computer is the most accurate way to judge its thirst for fuel.

WardsAuto editors have driven close to 800 cars over 20 years of Ward’s 10 Best Engines testing and written countless engine-related articles during that time. Still, some think my colleagues and I don’t know how to evaluate fuel efficiency.

When we say we use mileage data straight from the vehicle’s trip computer for our logs, there’s a sudden, disdainful look: “Really? I’ve always found my trip computer a bit optimistic when calculating my vehicle’s mileage,” they’ll say with a raised eyebrow and superior tone.

The suggestion is the time-honored method of calculating consumption by measuring gallons burned between fill-ups is the most accurate way to gauge because “everybody knows” automakers program their vehicles to give rosy reports on their secret gas-guzzling.

The engine and combustion experts we have spoken with over the years apparently are part of a vast collusion.

But there are problems with this conspiracy theory that defy logic. First, the rose-colored-glasses (RCG) software, if it exists, works poorly. During our test drives, almost all vehicle trip computers produce numbers that are well below the vehicle’s official Environmental Protection Agency rating. Sometimes the mileage results are downright ugly, thanks to occasional irrational exuberance on our part behind the wheel.

What’s more, the alleged RCG software is especially ineffective in hybrid- and plug-in electric vehicles. In our tests, diesel-powered vehicles usually are the only ones that exceed EPA estimates.

Shouldn’t RCG software be in full cloaking mode in HEVs and PHEVs to prevent the possibility of environmentalist-buyer remorse? Instead, there are lawsuits, especially from hybrid owners, upset about their vehicles not living up to advertised mileage claims, based on observed trip computer numbers. Shouldn’t secret software prevent lawsuits?

Secondly, as my mother used to tell me: “When you start telling lies, you eventually will get caught.”

Trip computers use fuel-efficiency calculations to count down how many miles of range are left in the tank. When does the RCG software start becoming more truthful to prevent drivers from getting stranded? Does it start breaking the bad news gently when it hits the half-empty mark, or keep lying until the engine sputters and dies while indicating 50 miles (80 km) of range is left?

Either way, wouldn’t this electronic chicanery result in millions upon millions of complaints about unreliable fuel gauges?

Last summer, when I raved about the 46.7 mpg (5.0 L/100 km) I clocked on a trip with the Chevy Cruze diesel Detroit Free Press auto critic Mark Phelan gave me the raised-eyebrow treatment when I told him trip computers are the most trustworthy source for vehicle fuel-economy information.

Much to my delight, he checked with a variety of experts at automakers and independent sources such as Brian West, head of the Oak Ridge National Laboratory’s fuels, engines and emissions research center and Frank Markus, technical director of Motor Trend magazine.

They all told him the same thing: Unless you want to spend millions of dollars on highly specialized measuring equipment, a vehicle’s trip computer is the most accurate way to judge its thirst.

“We need to know exactly how much fuel is being used by every stroke of every cylinder,” to meet emissions and fuel-economy standards, Roger Clark, senior manager of the General Motors Energy Center, told him. “We do a lot of precision measuring of the fuel injectors and exhaust treatment system.” The computer converts that data into fuel economy.

That kind of precision easily trumps Grandpa’s method, where temperature, sloping gas station parking lots and variations in the way gas pumps sense when a tank is full all can impact the outcome of fuel-economy measurements.


Discuss this Blog Entry 18

on Mar 14, 2014

Nonsense! A close family member (the one who likes to consistently tell me how wrong I am) insists that his Chevy 1/2-ton Z71 really gets mid to upper twenty's in mileage on the highway. Why? Because the instant readout tells him that the truck is doing something that the toy-car driving hippies/yuppies/city folk say doesn't happen.
However, I would like to see some mileage tests done on a gas 4-cyl whatever, in the hills of South Western Wisconsin when the temp is below zero. That'll wipe out an EPA rating really fast.

on Mar 17, 2014

Jef Von Q,
It sounds like you are responding to just the headline, not the story. Story is about calculated average fuel economy, not EPA estimates.

on Mar 17, 2014

While I love the diesels and how they often over-deliver performance and fuel economy, we have to remember that the EPA number we are told for their gas counterparts are often handicapped. The EPA test lab does not you pump gas (containing up to 10% eythanol), but uses 100% gasoline. That makes a huge difference in real world driving when it comes to fuel efficiency. Michigan has quite a few gas stations that offer 100% gasoline. Try it next time you fill up. You'll be surprised how much further you can go. That will explain why many people can't meet the EPA estimates on vehicles.

on Mar 18, 2014

I agree that the trip computer may be the most accurate way automakers can estimate fuel consumption economically. The way they calculate IS an estimate. I disagree that they are accurate. I did two 1000 mile fuel mileage tests with my Cruze Eco. I filled up at the same pump each time throughout the test and let the pump shut off on its own. Both times the trip computer was 2 MPG greater than the hand calculation.

on Mar 18, 2014

Not to throw another monkey wrench into the calculation, but within pump gas, there's a difference between the summer blend and the winter blend as well. All things are not equal when looking at the EPA estimates.

on Mar 19, 2014

I agree with Rocketman. My Kia Soul shows me an average of 35mpg, but when I calculate the average after filling up the tank, comparing miles traveled to the amount of fuel I put in, it comes up as 30mpg.

on Mar 19, 2014

My recent experience shows that the on-board displyed fuel economy is optimistic . For the first 10k miles of my new car I did not reset the trip computer fuel mileage calculation and recorded the volume of every fill-up. At the end of that 10k miles the vehicle displayed a fuel economy that was 5% better than the hand recorded values.
I investigated potential sources of error in my process but came up with no explanation for the discrepancy. There are, of course, just two elements to this calculation; volume and range. I looked at the for regulations concerning fuel pump measured volume accuracy (Note, with my volume recording method there is no error introduced due to automatic pump shutoff) and odometer accuracy requirements. Both are tightly regulated at state and federal levels. And after this study I am convinced that my calculation is closer to the truth than the on-board displayed value.

on Mar 19, 2014

The experts and the facts all say that hand-recorded values are not as accurate as the vehicle trip computer. A lower hand-recorded value than the trip computer underscores the error factor of the old method.

on Mar 19, 2014

I don't buy that. A volume error of 0.05 gallons in 15 and range error of 0.05 miles in 300, on average, accounts for only 0.5% error. Add the pump error and odometer error and you get nowhere near the 5% that I measured on the 1st 10k and have experienced over the subsequent 75k miles. And, as Erikgrad has mentioned, the vehicle calculation has always erred error on the side of optimistic; not distributed better and worse. I suppose I need to find that technical 'expert' and get the lowdown.

on Mar 24, 2014

Using a running calculation that spans many tanks and 10's of thousands of kilometers has always shown a systematic error in the FCD on every vehicle I've owned. If you fill up enough times you can predict your actual consumption as a percentage of the display in most cases.

As long as the sample size is great enough, dividing total fuel consumed by distance driven will be the most accurate. ( I would suggest a minimum of 5 tanks and a couple thousand kilometers as a good start).

You could start nit picking about how accurate fuel pumps and odometers are. But fuel pumps are certified by the government for commerce and odometers better be accurate or more lawsuits will result than fuel economy concerns would generate.

I don't know of any car that actually measures fuel flow or consumption directly, instead it is estimated based on injector pulse durations, fuel pressure maps and mass airflow rates etc. It is modeled relatively accurately but actually measuring the fuel consumed (what the gas pump does when you fill your tank) is inherently more accurate.

There is definitely a margin of error introduced with how full you fill you tank but if you use a running calculation that spans many tanks (see this error essentially disappears.

Many people love to only measure the good tanks/trips which leads to misleading average consumption claims, but again a running average that includes everything trumps this as well.

on Mar 19, 2014

On my Ecoboost Ford Edge, I can say with confidence that the trip computer is not accurate, and consistently indicates I use 10% less fuel than I actually have. Short of testing the pumps at the numerous gas stations I use, I cannot find a loophole that would cause it refill 10% more gas than the trip computer indicates I have used, 100% of the time. Instead, it should be fluctuating from one side of error to the other. I don't believe in an automaker conspiracy to fool me into thinking I am getting better mileage though - I just think the trip computer is fooled. This particular engine runs richer than perhaps it is supposed to (there is soot on the tailpipes), and the cold Michigan weather likely plays a role, as well...I have yet to drive the car in the warmth of summer. It may just be my model, and could just be my particular vehicle, so take it with an anecdotal grain of salt.

on Mar 19, 2014

Some of us have "prehistoric" vehicles that don't even have a trip computer, so we're forced to do the manual calculation.

on Mar 20, 2014

I like the convenience of trip computers, and I look to them in nearly 200 vehicles we drive as a staff each year at WardsAuto. Oddly enough, I see a lot of readings that are not very impressive, which brings into doubt the theory that trip computers are overly optimistic. But every so often, I do old-school hand calculations on the vehicles we drive. A few weeks ago, the BMW X5 trip computer read 16.8 mpg, while my calculation was 20.8 mpg. Last week, Toyota Highlander trip computer read 20.3 mpg while hand calculation indicated 19.9 mpg. This morning, Nissan Rogue trip computer reads 21.1 mpg but hand calculation is 21.6 mpg. There are so many factors that can affect both trip computer and hand calculations that truly gauging fuel consumption without expensive instruments can be a fruitless and maddening pursuit. Not a bad idea to do both.

on Mar 20, 2014

Or you could always just get a 4-cylinder diesel and not have to worry about your mileage or need to stop at the nearest station until you've clipped over 500 miles on the trip computer.

on Mar 21, 2014

I highly recommend the web where you can keep track of your MPG, fill-ups and over all fuel costs, the best it is free and all the calculations are done for you:‎

Second less publicized item that has significant effect on fuel economy at a minimal cost are better low friction lubricants such as:
The more fluids there are, like in Pick Ups, the bigger effect.
So vehicle that has multiple differentials, transmission and motor oil as well as power steering, will demonstrate bigger improvement than FWD car that has only Motor Oil and Trans Oil and all else (steering, etc.) are electric.

But unfortunately products like SynLube are available only in small volumes thus not practical for any OEM, but still individual vehicle owners can benefit.

COST PER MILE, should be the Measure that matters, and not any MPG, as at $$$ premium you can get Diesel, Hybrid or Pure EV, and at incremental cost that will buy lifetime of fuel for "low MPG" cars.

on Mar 22, 2014

My 2010 Toyota Prius Fuel Consumption Display (FCD) tends to estimate about 3 mpg more than my actual fuel efficiency computed at the pump (dividing miles driven by gallons pumped during a fill up). My Linear Logic Scangauge II xgauge Flv which monitors my fuel consumption indicates that the fuel gauge sensor cannot sense anything above 11.4 gallon even though the Prius technically can hold
11.9 gallons of gas and cannot see below 0.7 gallons albeit the Prius can still drink down to the last 0.2 gallon in the fuel tank depending on certain situations . However, depending on whether my Prius is on level ground or facing uphill or downhill - the sensor's fuel level rating changed. If the Prius is pointed uphill
then the fuel level seems more. If the Prius is pointed downhill then the fuel level seems less.

At 1.7 gallons - the Prius beeps and last bar of the fuel gauge flashes as a warning that I'm low on fuel -
and the miles before empty (MBE) estimation is usually about 30 miles ( but internally I have 2 gallons left). When Prius FCD MBE reaches zero - I still have about 1.5 gallons left - I heard that the MBE reaches zero early because designer wanted to avoid the gas tank from being completely empty.
Some nasty things can happen when the Prius runs out of gas...Don't get me wrong - I'm not complaining I am just staying that the instrumentation has its limitations but it is not a problem for a skilled driver. BTW my overall average over 3.5 years and 45000 miles is about 60 mpg with my MPG dropping in the winter and raising my summer ( for my fuel log look up Hyperdrive 1 on ). - much of my increased fuel efficiency has to do with how I drive my car ...I can tell you going above +65 mph really causes the Prius' fuel efficiency to drop like a rock.

on Feb 17, 2015

These is something that I haven't seen mentioned yet and I think Wardsauto's blanket statement leaves out an important fact or two. Checking our 2008 BMW X3 2008 Lexus IS250 has provided consistent and reparable results with similar inaccuracies with either vehicle. If all driving on a tank of fuel is in the city, each car shows a 1-2 mileage optimistic reading on the computer compared to manual calculations. This has been true for 7 years. That works out to 7-10%. But, on long drives such as I-81 with constant speeds, the variation is as little as 0.1 mpg which is meaningless. The only difference in computer accuracy is the effect of stop and go driving. The instantaneous readings are probably quite accurate but looking at the variance as you make small changes on the accelerator provides the clue to the problem. Fuel flow is continuous and can be shown by mapping flow over time. The total consumption then is the area under the curve. The smoother the flow line, the more accurate the calculation. The computer result could be made more accurate for stop and go but it depends on the small measurements of current consumption and really isn't that necessary but for Ward's to say it's accurate, not so, it just depends.

on Feb 17, 2015

Thanks for your thoughtful comments. I've been meaning to revisit this story.
After having spoken with a number of very smart, meticulous people, I believe a slight revision may be in order.
I still believe the car computer is more accurate than most drivers can calculate due to numerous variables such as temperature differences during fueling, angle of vehicle when being filled and many other factors.
Too many people insist their "old school" measurements are automatically more accurate than a vehicle's sophisticated electronics. That isn't the case.
However, I'm now convinced that super-meticulous drivers under specific, controlled circumstances may be able to calculate consumption more accurately than the trip computer, for some of the reasons you suggest.
The point of my story, though, is that for the typical driver, or for the WardsAuto panel of editors who have varied driving styles and testing circuits, the trip computer still will deliver more accurate readings by limiting variables.
Even so, we still do "old school" calculations in some cases for comparison.
Thanks again for weighing in. You and numerous other dedicated readers have helped evolve my thinking on this.

Please or Register to post comments.

What's Final Inspection?

WardsAuto editors share insights and observations on the global auto industry.


David E. Zoia

As Editorial Director, I oversee much of what goes into, enjoying a ringside seat that lets me observe up close just about every facet of the industry worldwide. I have covered the...

James M. Amend

James Amend is an associate editor at, covering day-to-day business and product news at General Motors. He also leads coverage of regulatory and environmental issues, as well as the...
Blog Archive
Follow Us

Sponsored Introduction Continue on to (or wait seconds) ×