Is GPS Speed More Accurate Than Your Speedometer?

Usually, yes. Your phone’s GPS measures how fast you actually move over the ground, while your car counts wheel turns and is built to read a little high. Most cars show 2 to 5% more than your true speed, so 60 mph on the dash is often 57 to 59 mph by GPS.

If your dashboard says 100 km/h and this page says 96, nothing is broken. That gap is designed in, and it is the single most common question people have about GPS speedometers.

The regulation

Under UNECE Regulation 39 — adopted across the EU, the UK, Australia, Japan and most other markets — a vehicle’s indicated speed must never be lower than the true speed, and must not exceed it by more than 10% plus 4 km/h. In the United States the requirements are looser, but manufacturers follow the same instinct.

Read that carefully: the permitted error is entirely one-sided. A speedometer that reads 5% low is illegal. One that reads 5% high is perfectly compliant. Every manufacturer therefore builds in a deliberate margin, because the cost of being caught under-reading is a recall and the cost of over-reading is a mildly annoyed customer.

In practice most cars indicate 2–5% high, which at motorway speeds is 3–6 km/h.

Why the margin has to exist

A conventional speedometer counts wheel or transmission rotations and multiplies by a fixed circumference. That circumference is not actually fixed:

• Tyre wear. A new tyre and one at the wear bars differ by around 1% in rolling radius.
• Pressure. An under-inflated tyre has a smaller effective radius and turns more often per kilometre.
• Aftermarket tyres. A different profile or width can shift the reading by several percent on its own.
• Load and temperature. Both change the radius slightly.

The manufacturer cannot know which of these apply to your car today, so they calibrate for the worst case that still never under-reads.

Why GPS has no such bias

A GPS reading does not care about your tyres. It measures the movement of the receiver through space, using the Doppler shift of the satellite signals. There is no calibration constant to drift, and no legal incentive to skew the answer in either direction. Its errors are random rather than systematic — exactly the property you want when the question is “how fast am I really going”.

So which one do I obey?

Your dashboard, for speed limits. Enforcement equipment is calibrated against true speed, and your speedometer over-reads, so obeying the dashboard keeps you comfortably legal. Deliberately driving to a GPS reading that sits above your indicated speed is how people collect tickets.

GPS, for everything else. Checking whether your speedometer is accurate after a tyre change, measuring a journey, timing acceleration, or settling an argument — use the satellite number.

Checking your own car’s error

On a quiet stretch of road, with a passenger operating the phone, hold a steady indicated 100 km/h and check how fast you are really going on the GPS speedometer. The difference, expressed as a percentage, is your car’s speedometer error. Repeat at 50 and 130 to see whether the error is proportional (usually) or constant (occasionally, on older mechanical units).

If the gap is more than about 10%, something is genuinely wrong — most likely the wrong tyre size fitted, or a sensor fault worth investigating.

Sources

• UNECE: Vehicle Regulations hosts Regulation No. 39, which requires the indicated speed never to be below the true speed and not to exceed it by more than 10% plus 4 km/h.
• Road Vehicles (Construction and Use) Regulations 1986, reg. 35 is the UK rule requiring a speedometer, and it accepts instruments approved under UNECE Regulation 39.
• W3C Geolocation API defines the coords.speed value this page reads, which comes from the receiver rather than from a wheel-speed sensor.