Why US Mileage Ratings Are So Inaccurate

Why does a car rated for 47mpg fall so far short? The Houston Chronicle features an article on just why EPA gas estimates can be so different from real-world drivers' experience at the pump (or in looking at the dashboard display), in particular for hybrid cars. From the article: "A geometric average of the FTP-75 and HFET results (with city driving weighted at 55 percent and highway driving weighted at 45 percent) produces a vehicle's CAFE fuel economy, which is then incorporated into a manufacturer's corporate average. CAFE is measured using these tests to the present day. In fact, this methodology will be 50 years old when it's used to gauge compliance with the forthcoming 54.5-mpg CAFE requirements in 2025. That kind of continuity is admirable in baseball, but not in transportation. These tests are irrelevant to contemporary real-world driving. For example, the maximum acceleration on either test is 3.3 mph per second. At that rate, it takes more than 18 seconds to hit 60 mph. Even in the horsepower-deprived 1970s, most people were driving harder than that. And the 60-mph maximum speed on the highway test does not accord with the 75-mph truth of today's interstate traffic."

Comments

consistency more important

[mondovoja]

Whether those numbers represent a real world mix of driving accurately really doesn't matter all that much, since fuel economy for other driving styles strongly correlates with fuel economy for the conditions that are actually measured. Long term consistency, on the other hand, matters a great deal for car buyers and for evaluating progress on reducing emissions and consumption.

Re:consistency more important

[swalve]

Yes, exactly. The CAFE ratings aren't meant to tell you what your personal MPG is going to be, they are meant to tell you how cars of a specific model year compare to each other. If you do 10% better in one car, you'll probably also do 10% better in the other one.

Re:consistency more important

[AthanasiusKircher]

If you do 10% better in one car, you'll probably also do 10% better in the other one.

That may be so, but miles per gallon is a misleading measurement on which to base the "10% better" calculation. 30MPG to 33MPG is *not* 10% better efficiency.

First, I would note that the GP said nothing about efficiency. The GP just said you would "do 10% better." MPG and gallons per 100 miles are just different measurements, useful for different types of calculations.

In some cases, MPG actually is a good measure, and 33 MPG is actually 10% better than 30 MPG.

For example, let's say that I'm not buying a car to use for my normal commutes, but I want a car to go on weekend road trips. I don't have a huge amount of money, so I have budgeted X amount of dollars for my fuel costs for my trips. Buying a car with 30 MPG instead of 20 MPG will actually get me 50% farther on my trips, which may mean I can take more of them or visit places that are farther away. In that sense, the 30 MPG car is 50% better than the 20 MPG car.

Similarly, let's say I'm still on a tight budget and looking for a new/different job. How far I can go on my potential commute might be based on the amount I have budgeted for gas. Again, in this case, a 30 MPG car will allow me to to have a 50% greater commute distance.

Is this a measure of greater "fuel efficiency"? No, not really. But, in cases where your budget or gas is constrained, MPG could actually be a measure of how much "better" a car could be for you.

If they switched to a burn rate measurement, like L/100km (that the rest of the world uses), or even Gal./100mi, then you actually could do the math in your head for how much more or less efficient the vehicle is.

I agree that this would be a better way to visualize fuel economy. But again, except for people who are really worried about the environment (not a bad thing), for most people the decision is going to be about practicality.

Your gal./100mi measurement would be great as a practical measure for someone who has fixed distances to travel rather than a fixed budget for gas. For people who use their cars primarily for commuting rather than for occasional road trips, they may just need to travel X miles per week. In such a case, a gal./100mi measurement will give an intuitive sense of what's "better" and by how much.

You're probably right that the gal./100mi system would be a better comparison for many (probably most) people.

But lots of people in the world rely on public transport for commuting and primarily use their cars for occasional tasks. And lots of people in the world have significant constraints on the amount of money they can afford to pay for gas. Those are the limiting factors for them, not a constant commute of X miles. For them, MPG is actually a more intuitive measurement of whether a car is 10% "better" from a practical standpoint.

Re:consistency more important

[GrumpySteen]

If they switched to a burn rate measurement, like L/100km (that the rest of the world uses), or even Gal./100mi, then you actually could do the math in your head for how much more or less efficient the vehicle is

Let's take a closer look at your point:

2L/100km is obviously twice as efficient as 4L/100km since 4 is twice as large as 2. Easy enough to do in your head.

50km/L is obviously twice as efficient as 25km/L since 50 is twice as large as 25. Just as easy since it's the exact same math.

The problem with your idea is that you read that a 10 mpg increase in fuel efficiency doesn't represent the same percentage of increase when it's applied to different starting mpg figures (i.e. from 30mpg it's a 33% increase, but it's a 50% increase if you start with 20mpg). And that's true, but you're assuming it affects all calculations using mpg figures. The L/km measure behaves in exactly the same non-linear fashion. A 1L/100km increase in efficiency from from 3L/100km is a 33% increase but it's a 50% increase from 2L/100km. Once again, the math is exactly the same.

tl;dr L/km isn't particularly better than mpg. You just suck at doing math that you haven't practiced, so you think it's harder.

Re:consistency more important

[nabsltd]

I think that over a certain speed, fuel economy goes way down.

The smaller the engine, the more you will see this effect. This is because it has to work much harder (relatively) to move the car faster.

For a larger engine on the same car, you generally get worse gas mileage at lower speeds because the idle fuel usage is higher. But, as you go faster, the larger engine doesn't work much harder, so mileage doesn't drop as much. After a certain speed, a larger engine will be more efficient on the same car.

This is why turbochargers are now all the rage. You get the advantage of the lower idle consumption of the smaller engine, while still having the acceleration you need for emergencies and the better mileage at higher speeds.

Your mileage may vary...

[dtjohnson]

No one cares what the testing procedure is as long as everyone does the same test and it's repeatable. The purpose of the test is to provide a method for consumers to compare different models with respect to their fuel economy, not to provide a precise prediction of exactly what the buyer's fuel economy will be. Everyone drives differently. People warm their car up in the driveway, fill it up with heavy weight, carry lots of passengers, do a lot of long-distance driving, tow trailers, drive up and down hills, ride their brakes, accelerate briskly to beat their neighbor, drive at high altitudes, drive in cold weather, or whatever. Even more significantly, the energy content of 'gasoline' varies widely depending on how much ethanol it has (more is less) and what its boiling point range is. Just do the same test and do it in a way that someone else could repeat the test the same way and get the same result. That's all we need rather it's a 50 year old test or not.

Re:Your mileage may vary...

[AmiMoJo]

The problem is that manufacturers tune their cars to do well in these tests at the expense of efficiency in more realistic conditions. Nobody accelerates at 3mph/s but cars are optimized for that because that's the test that is administered. If they made the test more realistic everyone would benefit.

Re:Drive conservatively!

[Totenglocke]

Hey, as long as you stay in the right lane when driving like a granny, I'm totally fine with it. It's when people pull that crap in the middle or left lane that makes me want to Hulk out.

Re:Drive conservatively!

[FlatEric521]

I agree. On my car with a built-in MPG screen, if I drive well, I get about 26.5 to 27 MPG on a car that's rated 23/31. Driving through town poorly, my wife gets about 23. Driving straight through to Vegas (from SoCal), I've gotten 32. And this was on the old "inaccurate" EPA scale. Seems pretty accurate to me.

I would recommend you consider double checking your car's trip computer calculation against the tripometer and gas pump readout method of manually calculating miles per gallon. I bought a Hyundai with a trip computer that includes Avg. MPG as one of its readouts. Prior to owning that car I had gotten into the habit of resetting the tripometer on my car at every fuel up after writing down the miles from the tripometer and gallons from the pump readout (then just divide the miles travelled vs gallons to refuel the tank). I kept up that habit after getting the Hyundai and found that the manual calculation method consistently reports 2-4 MPG lower than the trip computer. If your trip computer is anything as optimistic as mine, then you may actually be getting less than you think.

Why mileage ratings are inaccurate

[emaname]

I worked for just over 2 years in a wind tunnel for a company that manufactured cooling equipment (eg, radiators, oil coolers, A/C condensers and evaporators). We tested products for a variety of manufacturers which meant a wide variety of equipment; ie, compressors, farm tractors, semi tractors, passenger cars, and on one occasion a small city bus to be used in Miami, Florida.

We had a reputation for maintaining a very stable, controlled environment (air flow, heat load, dynamometer load, and positioning of thermocouples for sampling temperatures) and consequently consistent test results.

Now in the interest of full disclosure, this was in the early 70's. But at that time, that's also where the manufacturer's typically got their mileage estimates.

I think this might also be the era from where we get the expression "Your Mileage May Vary" (aka YMMV). I think they included this disclaimer in car ads in an attempt to comply with the "truth in advertising" laws (remember those?).

Clearly nobody can drive a vehicle in a manner as controlled as that.

So if the manufacturers are still getting their mileage results from a wind tunnel test, forget it. You'll never match those results especially if you live in a large metropolitan area (where it's not uncommon to sit idling in traffic) or you live in a mountainous area or where you have really cold weather.

There are several really good comments here with additional insight as to why mileage can vary drastically from the manufacturer's estimate; type of fuel mix, for one.

So remember, when you're buying a car and read those mileage estimates, YMMV.

Re:Drive conservatively!

[danbert8]

Incorrect. The correct driving lane is the rightmost lane you can occupy. If you are going 100mph you should still be in the right lane if you aren't passing. Idiots like you cruising along in the middle lane are why I end up passing you on the right. If you are in the correct lane you will NEVER be passed on the right.

Re:Speeds

[beelsebob]

Where I live the posted limits are 75 mph. Very few people here drive that slow unless the radar detector is going off or there is a cop actually in view.

The reason I know this is because I *do* drive the posted limit, and I am *constantly* being passed.

Arguing semantics But actually, that isn't valid reasoning. Even if you're constantly being passed, that's not a proof that the number of people passing you is greater than the number of people matching your speed exactly.