At a steady speed with a clear view of the sky, a phone’s GPS speed is typically within a couple of km/h (about a mile per hour) of your true ground speed, and often closer to the truth than a car’s speedometer, which is built to read high. It’s weaker when your speed is changing quickly, because it trails by a second or two, at very low speeds, and anywhere the sky is blocked: tunnels, dense downtowns, heavy tree cover. Hills barely matter. You can check your own phone in ten minutes with a known distance and a stopwatch.
A phone is good, not lab-grade. Dedicated GNSS speed meters used for vehicle testing update many times a second and do better, so treat tighter accuracy claims for phone apps with suspicion.
How does a phone measure speed? #
Most GPS receivers calculate velocity from the Doppler shift of the satellite signals, the tiny change in frequency caused by your movement. That’s steadier than working speed out from one position to the next. Phones typically hand apps a new reading about once a second, and apps smooth several readings so the number doesn’t flicker.
Speedometer GPS is a typical example of how an app handles this. It uses the phone’s own speed reading whenever the phone provides one, shows a rolling average of the last five one-second readings, and doesn’t use any fix with an accuracy worse than ±20 meters. If no valid fix arrives for five seconds, it drops to 0 and says SEARCHING rather than freezing on the last number.
When is GPS speed accurate, and when isn’t it? #
| Situation | How accurate | Why |
|---|---|---|
| Steady speed, open sky | Very good, typically within a couple of km/h | Doppler speed is steady |
| Hard acceleration or braking | Trails by a second or two | Smoothing; see GPS speed lag |
| Walking pace or slower | Noisier | Small errors are large compared with your speed |
| Among tall buildings | Jumpy | Reflected signals |
| Heavy tree cover | Weaker | Leaves absorb signal |
| Tunnels and garages | No reading | No satellites |
| Behind a metal-coated windshield | Weaker fix | Some heat-reflecting glass blocks GPS |
| Hills | Barely affected | See below |
Do hills affect GPS speed? #
Hardly at all. GPS speed apps generally report your horizontal speed, while you’re moving along a slope, so on a hill the true speed along the road is a little higher. The difference is tiny: on a 10% grade, speed along the slope is only about 0.5% higher than horizontal speed, which is 0.3 mph at 60 mph. On a steep 6% highway grade it’s under 0.2%.
So a hill can’t explain a 2 to 5 mph gap between your phone and your dashboard, and speed apps don’t need a “hill calibration” setting; none of the real ones have one. If the numbers disagree more on hills, look at the likelier causes: cruise control letting your speed drift up and down on the grade, a weak fix in a valley or under trees, or a speedometer that reads high everywhere and simply gets noticed on the climb.
A test you can run in a car #
Do this with a passenger reading the phone, or read the results after you’ve parked. Don’t look at the phone while you drive.
- Choose a straight highway with open sky, light traffic and mile or kilometer markers, where holding a steady speed with cruise control is legal and safe.
- Set up before you move. Mount the phone, open the app and wait until the accuracy indicator is good; in Speedometer GPS that’s a green EXCELLENT (within 5 m) or GOOD (within 10 m) pill. Press Start.
- Hold a steady speed and have your passenger time several markers with a stopwatch. True speed in mph is 3,600 divided by the seconds per mile; for kilometers, it’s 3,600 divided by the seconds per kilometer, in km/h.
- Compare. Your passenger notes the GPS speed during the timed stretch, or you save the trip afterwards and read the steady section off its speed chart.
Not every highway marker is placed precisely, so time several in a row and average them. If the phone agrees with the stopwatch speed to within a mile or two per hour, it’s doing its job, and your dashboard is probably showing a few percent more than both.
A test you can run on a bike #
- Find a measured stretch, such as a greenway or rail trail with kilometer or mile posts, or a segment you’ve already measured with a calibrated wheel sensor.
- Stop at the first marker, wait for a good fix and press Start.
- Ride to the second marker at a steady effort, stop, and press Stop. Time the ride with a stopwatch or note the trip time.
- Compare the average speed with the known distance divided by your time. Speed is the fair test, because it comes straight from the receiver.
If you also compare distance, ride above about 20 km/h (12 mph). Speedometer GPS ignores steps of less than 5 meters between one-second fixes to keep drift out of trips, so below about 18 km/h (11 mph) its distance reads short even when its speed is right. At a brisk pace on a one-kilometer stretch, a difference of 10 to 20 meters is 1 to 2%, which is normal. For more on how GPS compares with a wheel sensor on a bike, see bike speed sensor vs GPS.
Reading your results #
- Within a couple of percent on distance, and a couple of km/h on speed: normal for a phone.
- Off in the same direction on every run: check that you started and stopped at exactly the markers, and that the fix was good.
- Big swings between runs: the fix is weak. Move the phone somewhere with more sky.
Why GPS often beats the dashboard #
Car speedometers are required never to read low. Under UN Regulation No. 39, the indicated speed must be at least your true speed and may be up to 10% plus 4 km/h above it, and worn tires push it a little higher still. So if your dashboard says 70 mph and GPS says 67 on a steady road, it’s normally the dashboard that’s off. Our guide to checking and calibrating a speedometer with GPS shows how to measure your car’s own error.
Once you know your phone is reliable, you can use it as your reference. In Speedometer GPS, a speed alarm set to the posted limit warns you at your true speed, and each saved trip keeps a speed chart you can check after you park.
Frequently asked questions #
How accurate is GPS speed on a phone? #
At a steady speed with open sky, typically within a couple of km/h, or about 1 mph, of your true ground speed. It’s less accurate during hard acceleration, at walking pace, between tall buildings and under trees, and it doesn’t work in tunnels.
Is GPS speed more accurate than a car speedometer? #
Usually, at a steady speed with a good fix. Car speedometers are designed to read at or above your true speed and commonly show a few percent high, while GPS measures actual ground speed. The dashboard responds faster to sudden changes, though.
Do hills make GPS speed inaccurate? #
Not in any way you’d notice. Even on a 10% grade, the difference between speed along the slope and horizontal GPS speed is about 0.5%, or 0.3 mph at 60 mph. Bigger gaps on hills come from changing speed, weak signal in valleys, or a speedometer that reads high.
Why is my GPS speed different from my speedometer? #
Mostly because speedometers are built to read high, typically by a few percent, and worn tires add a little more. GPS also trails by a second or two while you’re speeding up or slowing down. At a steady speed with a good fix, the GPS number is normally the closer one.