How EPA Fuel Economy Ratings Are Measured
Last updated July 24, 2026
The number on a car's window sticker comes from a standardized laboratory test, not a road trip. Here is how EPA actually arrives at that figure, and why your own results will still vary.
Tested in a lab, not on the road
Every fuel economy number on a window sticker, or on a vehicle page on this site, comes from a laboratory test rather than a drive on public roads. The vehicle sits on a chassis dynamometer, a set of rollers under the drive wheels that simulate road load, while a technician follows a fixed speed and acceleration script for each test.
Running every vehicle through the identical script, indoors, with the same simulated conditions, is what makes fuel economy numbers comparable from one model to another in a way a real road test never could be, since real roads vary in traffic, weather, and terrain every time you drive them.
The city and highway test cycles
The city figure comes from a cycle called the FTP 75, built from a real trip recorded in Los Angeles rush hour traffic decades ago. It runs about 11 miles, averages roughly 20 miles per hour, and includes frequent stops and starts.
The highway figure comes from the HWFET cycle, about 10 miles with an average speed near 48 miles per hour and no stops at all. The combined rating most shoppers see is a weighted average of the two, typically 55 percent city and 45 percent highway, meant to represent a typical mix of driving rather than either extreme on its own.
Why the numbers were adjusted in 2008
For decades, EPA stickers were based only on those two cycles, and drivers routinely found they could not match the number in ordinary driving. Starting with the 2008 model year, EPA began applying downward adjustment factors to the raw dynamometer results, roughly 10 percent off the city number and 22 percent off the highway number, and added three more test cycles to the process: a high speed and hard acceleration cycle, a cycle with the air conditioning running, and a cold temperature version of the city cycle.
Together these five cycles produce a rating meant to track real world results far more closely than the original two cycle test did, though it is still a laboratory estimate rather than a personal guarantee.
Why your own number will still differ
Even the corrected rating is an average across a standardized script, and no individual driver matches that script exactly. Aggressive acceleration and braking, sustained speeds above the test cycle's averages, cold weather, running the air conditioner or heater harder than the test assumes, carrying extra passengers or cargo, underinflated tires, and even the ethanol content of the fuel in your tank can all move real world results away from the rating, in either direction.
Treat the EPA number as a standardized way to compare vehicles against each other, not as a promise about what your own dashboard will read.
What else is on the window sticker
The city, highway, and combined MPG figures are the headline numbers, but the same fuel economy label also carries an annual fuel cost estimate, a five year fuel cost comparison against an average new vehicle, and an estimated cost to drive 25 miles. These are all derived from the same underlying test results, run through EPA's assumed 15,000 miles a year of driving and a national average fuel price that gets updated periodically, which is why the annual cost figure you see quoted for a specific vehicle can shift over time even though its MPG rating never changes.
Newer stickers also carry a smog rating and, for gasoline and diesel vehicles, a greenhouse gas rating, both on a 1 to 10 scale, alongside the fuel economy numbers. These environmental ratings are calculated from the same test data as the MPG figures, since fuel burned and CO2 emitted are directly proportional to each other for any given fuel.
Measuring your own real world number
If you want a number that reflects your own driving rather than the standardized test, the tank fill method is the most reliable way to get one. Fill your tank completely and reset your trip odometer, drive normally until you need fuel again, then fill up again and record exactly how much fuel it took. Dividing the miles driven by the gallons used gives you a genuine real world figure, and doing this over two or three fill ups smooths out any single tank that was unusually short or unusually highway heavy.
This is worth doing even if your car has a built in fuel economy display, since those dashboard estimates are calculated on the fly from sensor data and commonly read 5 to 15% more optimistic than what a tank fill calculation shows. A tank fill number is also the right input to use in a fuel cost calculator if you want the most accurate possible cost estimate, rather than the EPA rating or the dashboard's own estimate.
Frequently asked questions
Yes. Every vehicle sold in the United States runs the same standardized city, highway, and supplemental test cycles on a dynamometer, which is what makes EPA ratings a fair way to compare one model against another.
The most common reasons are driving faster than the test cycle's average speeds, frequent hard acceleration, cold weather, and short trips that do not give the engine time to reach its most efficient temperature. All of these push real world results below the lab estimate.
MPGe converts an electric vehicle's energy use into a figure comparable to MPG so gas and electric vehicles can be compared side by side. See our MPG versus MPGe guide below for the full explanation.
EVs go through an analogous dynamometer process measuring electricity consumption instead of fuel burned, using the same underlying city and highway driving cycles, then that energy use is converted into the MPGe figure shown on an EV's window sticker using EPA's fixed gasoline energy equivalence constant.
The core five cycle test methodology has been in place since the 2008 model year adjustment described above, and has proven durable, though EPA periodically reviews and can revise details like the assumed national average fuel price used in annual cost estimates without changing the underlying MPG test itself.
Yes, commonly. Weight, aerodynamics, gearing, tire choice, and all wheel drive versus front wheel drive all affect the result even when the engine is identical, which is why EPA tests each individual trim and configuration rather than assuming one rating applies across an entire model line.