By on July 24, 2014


As reported earlier, the Environmental Protection Agency is planning to put automakers through real-world testing of their MPG claims. However, the agency is asking this be done on the track instead of the highway.

Edmunds reports the plans are in the “consideration phase,” with the aim of bringing claims in line with what drivers experience on the road, according to an unnamed EPA spokeswoman:

EPA is considering requiring automakers to perform supplemental test-track audits of production vehicles to validate the values for aerodynamic drag and tire friction, which are important data inputs for our laboratory fuel economy testing. Augmenting EPA’s existing pre-production procedures with post-production audits of real-world factors will help further ensure that the data used in EPA labels accurately reflect the vehicles consumers find on dealer lots.

She goes on to clarify that the testing should occur on the track, and that the agency is “not considering creating public roadway test procedures” to replace the results currently gained in the lab.

Get the latest TTAC e-Newsletter!

25 Comments on “EPA In “Consideration Phase” Of Real-World Fuel Economy Testing...”

  • avatar

    Just stop doing it entirely, please.

  • avatar

    I can see one of two affects.

    1)It ends the rediculous shift to overly complex micro engines, in place of more efficient in the real world N/A engines.
    2) It results in all fuel economy numbers being redone at lower numbers, thus causing CAFE to go in overdrive mode for the automakers.

    …. Or nothing changes at all…. That’s probably the most likely scenario, they’ll just waste a couple billion and say we’re now all safer knowing current MPG figures are correct.

    • 0 avatar

      EPA wont be changing how CAFE measures. This is for consumers.

      CAFE already uses the old EPA figures that you haven’t seen on window stickers for about 10 years now.

  • avatar

    modern automatic transmissions are designed to game the EPA testing, not acutally be efficient. Looking forward to seeing how well they do agaist manual transmissions in ‘real world’ testing.

    • 0 avatar

      All I can hear is the sound of Ford executives shutting themselves.

    • 0 avatar

      It’s not just the automatics. Take the new 6mt in the Honda Fit that has the same top gear ratio as the old 5mt. More shifting for the same crappy real-world highway mileage, but I bet it comes out better on the EPA test.

      • 0 avatar

        That is more likely due to people not wanting to downshift their MT pulling moderate hills on the interstate rather than gaming the EPA test. On ATs, it isn’t as unpleasant to downshift when already traveling at 65 or 70mph as it is in a low power, low torque MT. Having owned a 1993 Impreza with 110hp (on a good day), AWD, and a 5MT, I can attest that it is pretty unpleasant to rev match to get the car to 5000RPM to pull the Cheat Mountain hill on I68.

        • 0 avatar

          I’ve heard this logic, but I still find it lacking.

          Put a very tall overdrive gear in as 6th, taller than the existing top gear. If people want to drive in a shorter gear with no downshifting, they can leave it in 5th. Problem solved.

          • 0 avatar

            It is what it is. Automakers have focus groups to tell them how the car should drive or what is important to the buyers. The fact that so many stick shifts with tiny engines have aggressive cruise gearing means that the customer is asking for that. It makes that tiny engine feel like a less tiny engine.

            If it was for the EPA test ratings, you’d see MT gearing much more like AT gearing. Gaming the EPA test would be setting your gear ratios to provide just enough power to maintain speed in top gear through the EPA test cycle. Having the car be responsive in top gear like the above mentioned Fit complaint is pretty much the opposite of gaming the EPA test. Based on that, I can’t think there would be any other reason that customer preference to do MT gearing the way they do on the Fit and similar cars.

          • 0 avatar

            @Quentin: Pretty much spot on. Manufacturers would love to put in a much taller top gear for better cruising, but people don’t like the feeling that the car is “underpowered” and needs to be downshifted to keep up with traffic, so they need to keep the little cars buzzing along at over 3k rpm at 65 to make sure no one complains.

            With an automatic, there’s little issue, the car does the downshifting for you.

        • 0 avatar

          We are very different indeed…

          Revving the living piss out of a tiny engine just to keep moving, is one of driving life’s few remaining pleasures in my book. As long as the car has a manual. With an auto, even one as good as the latest 7/8 speeds, the constant 6-7-8 shifting over even the smallest hill, annoys me enough to just put the tranny in sport mode and negate any possible fuel savings.

  • avatar

    This subject is much more complex than people realize. The EPA test, done on a dynamometer under strict conditions in a lab, is significant because it provides a standard that all manufacturers can repeat time and time again. There have been all kinds of adjustments to the test itself to try to make the results more consistent with people’s real-world experiences, but there still is the dreaded Your Mileage May Vary asterisk. In the real world, people rarely, if ever, drive the way the EPA test is designed and hence get upset when their new vehicle doesn’t exceed the ratings.

    So the EPA is considering how to add a real-world driving test to their measurements. This makes things much more complicated. How many miles have to be traveled? What ambient temperature? What altitude? What type of gasoline is used, winter oxygenated, or summer blend? How good is the test driver and how well do they follow the program of acceleration (at what rate?), deceleration, and time at various speeds? What is the condition of the road? How many changes of elevation are allowed? All of this stuff is controlled in the current test. Putting a car out on the road brings all this into play again.

    It’s engineering, yo.

    • 0 avatar

      I bet manufacturers actually have very good real-world data from their own internal testing, but I have no clue how you’d get them to honestly provide it.

  • avatar

    Track testing yields more consistent results than does highway testing. On the track, you don’t have to fight conflicting traffic. However, testing on a dynamometer in a closed building is still better.

    All along, the problem with the EPA’s results has been that they use an unrealistic duty cycle. Developing one that matches how people actually drive would be difficult. The only way I can think of is to instrument a large number of vehicles of all types to record the parameters needed to program the dynamometer test. This would be a monumental project.

    • 0 avatar

      I disagree. You could collect that data from about a thousand cars and get a pretty good average. There isn’t enough grade and driving style variety for this to be distinctly outside the norm.

      The hard part isn’t coming up with an average road/load, it’s matching driving style to car. I guarantee an average BMW driver will get worse mileage in their car than the EPA would estimate if they’re basing the rating on driving habits of geriatric Buick drivers.

  • avatar


  • avatar
    heavy handle

    I’m surprised how many people who spend their days complaining about EPA mileage estimates don’t actually understand how it’s done.

    What they are talking about here is coast-down testing. Manufacturers use coast-down figures to estimate aero and mechanical drag, which is then added to EPA runs done on a dyno (because there’s no aero drag on a dyno, and some mechanical drag is also absent).

    In the past, manufacturers were allowed to estimate coast-down figures based on engineering models. The problem with this is that manufacturers would fudge the numbers, boosting their EPA ratings. Hyundai and Ford both got dinged for this, which resulted in huge fines and refunds to owners.

    Short version: the EPA just said “you can’t pull coast-down numbers out of your ass anymore. You have to measure them in the real world using real cars.”

    Will this cost American any money? No. Manufacturers already run these tests. If anything, it will save Americans money because they will have access to more accurate information.

    • 0 avatar

      I somehow read right over the track validation of the rolling and aero numbers. Validating those values makes a lot more sense than trying to come up with some sort of procedure that validates the actual EPA test cycles in the “real world”.

      IMO, my Prius v Five should have a lower EPA rating than the Prius v Two and Three trim packages. Wider tires on 17″ wheels that my Five comes with will definitely impact mileage over the narrow tread, light 15″ wheel/tire combo on the Two and Three. I knew that would likely be the case going in and the features of the Five trim were more important than the margin fuel savings of the Two/Three trims, so I’m not disappointed, but I could see some people expecting them to be the same. The same goes for offroad packages on trucks. Those offroad tires definitely impact fuel economy. Strangely enough, Toyota does put different numbers on the Camry Hybrid LE and XLE based on the wheel tire package.

    • 0 avatar

      I was going to post exactly this.

      Cameron–the Edmunds article is a poor explanation of the issue, and I would hope that TTAC would dive into the issue and break through the confusion of that article. The quote you picked to include is excellent, but I’d like to see more analysis:

      “EPA is considering requiring automakers to perform supplemental test-track audits of production vehicles to validate the VALUES FOR AERODYNAMIC DRAG AND TIRE FRICTION, which are important DATA INPUTS for our laboratory fuel economy testing. Augmenting EPA’s existing pre-production procedures with post-production AUDITS OF REAL-WORLD FACTORS will help further ensure that the data used in EPA labels accurately reflect the vehicles consumers find on dealer lots.” (emphasis added)

      This makes it clear that the EPA is *not* doing real-world mpg tests, nor are they forcing automakers to do different mpg tests–they are testing cars for characteristics to use in their existing procedures. Therefore, it makes perfect sense that there would be no public roadway tests, not to mention that public roadway tests can’t be controlled like track tests. Frankly, everyone who suggests public roadway tests doesn’t understand mpg tests.

      Comments about making sure the mpg ratings match what customers see is misleading because they suggest that they will use the real cars to get real mpg. Instead, it’s only about improving inputs into an equation so the answer isn’t bogus. It’s good that this report explicitly states that the EPA isn’t replacing their tests, but it deserves a bit more emphasis because that’s not what most people would read.

      However, the variance of hybrids may pose a problem. The discrepancy those cars see in EPA mpg & reported mpg suggests simply improving drag & friction parameters won’t reconcile the two. The EPA’s actions won’t make that problem go away.

      • 0 avatar

        Lots of great logic thought into the posts tonight.
        I agree that on road testing is complex, and still has the potential for “gaming” to occur.
        What I’d prefer to see is a better representative dyne testing, especially around the coast down loads. We (BB) all know about the famous, super smooth, with afternoon tail wind track in Spain used for developing coast down loads.
        My 2c worth.
        I’d like to see better real world representative track built, which uses high friction bitumen (representing the average American road), that manufacturers are forced to use, and for the drive pattern to be updated to represent something more realistic of our daily drives. Maybe more than just city/highway.

        • 0 avatar

          They already have 5 tests: City, Highway, High Speed (which has up to 8.46 mph/sec acceleration and a top speed of 80 mph), A/C, and Cold Temp. They Use the latter three to adjust the first two.

          What would make it more realistic to “our daily drives?” This is a big country and drives vary considerably.

  • avatar

    This would be interesting to see and hear from those who are still going to complain. I’ve been able to match or best the EPA numbers in every 4 cylinder car I’ve had, the 6s on the other hand…

    Then you have those like my brother who measures mileage based on tank. I was telling him the mileage estimates for the Buick he’s borrowing and he said, “this means nothing to me, I go based on how many tanks I’ll go through in a week.” I just shook my head and walked away.

    • 0 avatar

      I’ve noted the same – a four with row-your-own gears will almost always meet or exceed the EPA sticker, especially since they reduced all the estimates. If I got the EPA city rating I’d take my car in to find out what’s wrong with it.

      A six with an auto will usually do pretty close on the highway as long as you are gentle with it, but they just suck the gas away in the city.

  • avatar
    George B

    The low-hanging fruit of improving the accuracy of EPA window sticker fuel efficiency numbers are 1) accurately measuring the aerodynamic and tire drag as proposed and 2) including higher driving speeds in their test cycle. The big problem for the consumer is normal Americans think highway speed is about 70 mph, but the “highway” test cycle assumes lower speeds where aerodynamic drag is less of an issue. Combine the EPA cycle with European car regulations and American consumers get stuck with tall cars that have to punch an excessively big hole through the air. Lower and longer cars with proportionally smaller wheels could have better real-world highway fuel efficiency, but current rules and incentives discourage this shape of car.

  • avatar

    Cars are getting taller because, as with “regular” cars, people are starting to expect more and more space from “small” cars. Which means either they grow taller or longer or both.

    That tall and short cars are not as economical as longer and lower cars is not lost on automakers. Do note the most recent “mini” cars, such as the Mitsubishi Mirage and the Hyundai i10 (Europe and Asia only, sorry)… are both longer and lower than their competitors.

    The Mirage has a 0.27 – 0.28 Cd.

    For those of us doing economy testing, the EPA is actually way too conservative/pessimistic. Nobody who actually cares enough about economy to adjust their driving style for best economy (not to the point of driving at 40 mph on the highway, mind, but simply staying under the posted limits and not driving like a bat out of hell) gets below EPA.

    In the end, the real problem with the EPA is spotty auditing. Which this move hopes to rectify. If you’re an economical driver, you will consistently get x% better than EPA on cars which are rated properly. If you’re a “normal” driver, you will consistently get y% worse.

    All that’s necessary, really, is to ensure that everyone is using the same yardstick.

Read all comments

Recent Comments

  • stuki: When where you drive requires lots of speed changes: The combination of a rather slow shifting transmission,...
  • EBFlex: “ Kia still offers the ten-year warranty with the Telluride, which is enough to set off a mental alarm bell....
  • JMII: Imagine the payouts if BMW offered such a long warranty. I am half convinced that the Germans are so good at...
  • Astigmatism: That screamed to me “salesman for another brand.” Like, you walk into a Toyota dealership,...
  • ScarecrowRepair: That struck me too: “Kia still offers the ten-year warranty with the Telluride, which is...

New Car Research

Get a Free Dealer Quote

Who We Are

  • Adam Tonge
  • Bozi Tatarevic
  • Corey Lewis
  • Jo Borras
  • Mark Baruth
  • Ronnie Schreiber