On open roads, a phone mounted on your handlebars and a GPS bike computer read speed to about the same accuracy, because both measure it from the same satellites. The bike computer’s real advantages are elsewhere: it can pair with a wheel sensor that reacts instantly and works under trees and indoors, its screen is built for sunlight, its battery is built for long days, and it connects to power, cadence and heart-rate sensors. A phone is plenty for commuting, casual rides and knowing your speed. Structured training and long, fast rides favour a dedicated computer.
How do phones and bike computers measure speed? #
GPS speed. Phones and GPS bike computers both take speed mostly from the Doppler shift of the satellite signals, about once a second, then smooth it over a few seconds so it doesn’t flicker. It needs no setup, but it needs a view of the sky, and smoothing makes it trail sudden changes.
Wheel speed sensor. A small sensor on the hub (or a magnet and pickup on older setups) counts wheel turns, and the computer multiplies them by the circumference you entered. It reacts within a wheel turn and doesn’t care about the sky, but its accuracy depends entirely on that circumference. Most bike computers use the sensor’s speed when one is paired and fall back to GPS when not. Bike speed sensor vs GPS compares the two in detail.
Can the phone’s other sensors measure speed? #
No. The accelerometer, gyroscope, barometer and compass are good at other jobs, but none can hold a speed for more than a few seconds. To get speed from an accelerometer you add up every tiny change, and the errors add up too. A bias of just 0.05 m/s² becomes an error of about 11 km/h after one minute. Misjudge the phone’s tilt by one degree and a sliver of gravity looks like forward acceleration, about 6 km/h of error after ten seconds. Some phones blend motion data into their location to smooth brief gaps, but on a phone bouncing on handlebars that bridging lasts seconds. An app promising sensor “fusion” that keeps your speed accurate through a long wooded descent is promising more than the physics allows.
Phone vs GPS computer vs sensor at a glance #
| Phone on the bars | GPS bike computer | Bike computer + speed sensor | |
|---|---|---|---|
| Speed on open roads | Good | Good | Good, once calibrated |
| Under trees, in tunnels | Weakens or drops out | Weakens or drops out | Unaffected |
| Response to a sprint | A second or two behind | A second or two behind | Instant |
| Distance on twisty trails | Tends to cut corners | Tends to cut corners | Follows your actual path |
| Screen in bright sun | Readable, but brightness drains battery | Usually designed for it | Same |
| Battery on long rides | A few hours without a power bank | Usually much longer | Same |
| Indoor trainer | No | No | Yes, on the driven wheel |
| Power, cadence, heart rate | Depends on the app | Yes | Yes |
| Cost | Usually already owned | Extra | Extra |
Why do riders see different numbers on the same group ride? #
Everyone in the bunch is moving at the same speed, yet the numbers differ. Instant speed on two good devices usually agrees within 1 or 2 km/h. Averages can differ far more.
| Cause | What you see | Which number it affects |
|---|---|---|
| Different smoothing | One device reacts faster to surges | Instant speed, max speed |
| Recording interval | Devices recording every few seconds cut corners | Distance, average |
| Auto-pause rules | Devices stop counting at different speeds | Average, moving time |
| Moving vs elapsed average | One counts the café stop, one doesn’t | Average speed |
| Wheel sensor vs GPS | Sensor unaffected by trees, but depends on wheel size | Speed, distance |
| Phone placement | A phone in a back pocket sees less sky | Everything, under trees or buildings |
| Start and stop points | Someone started at the café, not the car park | Distance, average |
Averages are where the big gaps come from. Take a 60 km ride with 2 hours of riding and a 30-minute café stop: the moving average is 30 km/h and the elapsed average is 24 km/h. Neither device is wrong.
Max speed depends on smoothing. A device that averages over five seconds records a lower top speed on a short sprint than one that averages over one second. A device that smooths too little can record a false maximum from a single glitchy reading under trees.
Drafting and wind don’t change GPS speed. GPS measures speed over the ground, whether you’re on the front in a headwind or tucked in fifth wheel.
What about distance? #
Over a whole ride on open roads, GPS and a calibrated sensor usually land close. On twisty descents and singletrack, GPS cuts corners between fixes and reads short, while a sensor follows every metre the wheel rolls. Bike computer vs GPS distance explains the gaps and how to calibrate a sensor.
Elevation is a separate question. Many bike computers and many phones have a barometric altimeter, which handles climbs much better than GPS altitude alone. Whether an app uses the phone’s barometer is up to the app, so treat GPS-only elevation gain as rough.
When is a phone enough? #
A phone is enough when you want speed and distance, not analysis of every pedal stroke:
- Commuting and errands.
- Casual and social rides.
- E-bikes, scooters and older bikes with no speedometer.
- Checking whether a bike computer’s wheel-size setting is right.
Speedometer GPS is built for that job. Its dial auto-ranges, so on a bike you get a fine 0 to 60 km/h (0 to 40 mph) scale, and the number is the average of the last five one-second readings, steady enough to read on rough roads. The accuracy indicator under it shows the fix quality and radius, so you can tell a real surge from a weak signal. Save a ride and you keep distance, top speed, average speed while moving, moving time, the route and a speed chart, with GPX export, and it works without mobile data.
Its limits are real. It can’t pair with a speed sensor, power meter or heart-rate strap, it has no auto-pause, and it has to stay open on screen to record. Its trip distance also under-counts below about 18 km/h (11 mph), because it ignores steps of less than 5 metres between one-second fixes to keep a parked phone from adding distance. On slow climbs, in traffic and on easy spins, that costs real distance. Live speed, top speed and average speed aren’t affected, and average speed × moving time gives a fair estimate. If whole-ride distance matters, use a sensor or cycling app for distance and the phone for speed.
How to set up a phone as a bike speedometer #
Do all of this before you roll. Nothing should need a touch once you’re moving.
- Use a solid stem or bar mount, not a jersey pocket. Turn the bars to full lock both ways to check nothing hits. Check the mount before every ride; on a bike the main risk is a mount that lets go on a pothole.
- Allow precise location while using the app, and stand outside until the accuracy indicator shows a good fix. In the open that takes seconds.
- Pick km/h or mph and set any speed alarm.
- Press Start as you roll off and leave the app on screen. Keep-screen-on is on by default, because a locked phone stops recording.
- Press Stop at a café stop to pause, and Start to carry on. At home, press Save.
- Glance only on straight, clear stretches, never in a corner, a busy junction or a tight paceline. Silence notifications so nothing covers the readout.
If no good fix arrives for five seconds, under a bridge or dense trees, the speed drops to 0 and the app shows it’s searching until the signal returns.
Using the speed alarm on a bike #
The speed alarm isn’t only for cars. Set a ceiling in 5 km/h or 5 mph steps and the readout turns red and the phone buzzes and beeps when you pass it, repeating every 10 seconds until you slow. Riders use it on shared paths with posted limits, to keep an e-bike under a set speed, or as a reminder on descents they don’t know. You hear it without looking down, which is the point.
Battery, heat and cold #
A fix every second plus a lit screen is steady work. On rides over a couple of hours, run a short cable from a power bank in a top-tube bag. Cold shortens battery life temporarily, and a phone in full sun on the bars can get hot enough to dim its screen. Our battery drain guide has more.
Frequently asked questions #
Is a phone accurate enough for cycling speed? #
Yes, on open roads with the phone on the bars and a good fix. At a steady pace it’s typically within a couple of km/h. It lags a second or two in sprints and struggles under heavy tree cover, where a wheel sensor does better.
Why does my friend’s bike computer show a different speed from mine? #
Devices take their own fixes with small independent errors and smooth them differently, so a 1 or 2 km/h gap is normal. If either of you uses a wheel sensor, check the circumference setting; a wrong value makes one device consistently high or low.
Why is my average speed lower than everyone else’s? #
Usually because your device shows an elapsed average, including stops, while theirs shows a moving average. Starting to record earlier, such as on the slow roll to the meeting point, also pulls it down.
Does GPS speed work on an indoor trainer? #
No. You aren’t moving over the ground, so GPS reads zero. Indoors you need a wheel speed sensor on the driven wheel or a smart trainer that reports speed.
Does average speed include stops at traffic lights? #
In Speedometer GPS, no. Average speed counts only while you’re moving, and the saved ride shows moving time separately from total time, so red lights and café stops don’t drag it down.