A wheel-sensor bike computer multiplies wheel turns by the circumference you entered; a GPS app adds up the straight lines between position fixes. On open roads with a correctly measured wheel, the two usually land within a percent or two. Bigger gaps nearly always come from a wrong circumference setting, GPS cutting corners on twisty routes, or GPS losing its fix under trees. Hills make almost no difference, despite what you may have read. Once measured, a calibrated sensor is usually the more accurate of the two, typically within about half a percent.
Why don’t the two distances match? #
| Cause | Effect on the wheel sensor | Effect on GPS |
|---|---|---|
| Circumference set wrong | The same percentage error on every ride | None |
| Tyre change, say 25 mm to 28 mm | About 1.5% off until you update the setting | None |
| Low pressure or a heavy load | Tyre squashes, circumference shrinks slightly | None |
| Tight switchbacks and singletrack | Counts every wiggle | Cuts corners, reads short |
| Slow climbs | Unaffected | Fixes wobble; apps add zigzags or filter out real distance |
| Tree cover, canyons, city centres | Unaffected | The fix wanders; distance can drop out or wobble |
| Magnet slipped, sensor battery flat | Misses turns, reads short | None |
| Wheel slip on mud or loose climbs | Counts turns that didn’t move you | None |
| Hills | Measures along the slope | Adds up horizontal distance; the gap is tiny |
Do hills really add distance? #
Barely. A wheel rolls along the slope, while GPS apps normally add up horizontal distance, as if the road were drawn on a flat map. Hills are much longer than they are tall, so the slope is hardly longer than its base:
| Grade | Example | Extra distance along the slope |
|---|---|---|
| 5% | A long highway climb | 0.12% |
| 8% | A stiff road climb | 0.32% |
| 10% | A steep road climb | 0.50% |
| 15% | A very steep lane | 1.1% |
| 20% | A wall most riders walk | 2.0% |
| 25% | The steepest paved streets | 3.1% |
Put that on a real ride. Say you ride 50 km with 15 km uphill at 7%, 15 km back down and 20 km of flat. The slopes add about 70 metres in total, 0.15% of the ride, which a display rounding to a tenth of a kilometre won’t even show. The old claim that a hill turns a mile into 1.2 miles is a myth: a 500-foot climb over a horizontal mile adds less than 25 feet.
So if your sensor shows 3% or 5% more than GPS after a hilly ride, the slope isn’t the reason. Hilly rides tend to be twistier, slower and more wooded, and that’s where GPS loses distance.
Switchbacks and corner-cutting #
Phones and bike computers take a fix about once a second and join them with straight lines. At 20 km/h that’s a point every 5.5 metres or so. On a straight road the shortcut is invisible. On hairpins and singletrack, the straight lines skip the inside of every bend. Devices that record every few seconds to save battery cut far more; why GPS distance varies by sampling rate works through the numbers.
Slow climbs and tree cover #
At 8 km/h you move about two metres a second, less than the error in each GPS fix. Some apps count that wobble as extra distance; others ignore small movements to avoid counting drift and end up missing real distance. Leaves, especially wet ones, weaken signals, and steep valley walls hide part of the sky. Climbs in hill country tend to have all three problems at once.
Which number is right? #
| Question | Better choice | Why |
|---|---|---|
| Total distance on open roads | Either | A calibrated sensor and GPS should agree closely |
| Distance on switchbacks and twisty trails | Wheel sensor | GPS cuts corners |
| Distance on slow climbs under trees | Wheel sensor | GPS fixes are at their weakest |
| A route map to share | GPS | A wheel sensor doesn’t know where you went |
| Catching a wrong wheel-size setting | GPS, on an open road | It needs no calibration |
For training logs, consistency matters more than the last percent. Pick one method and stick with it, so this month compares fairly with last month.
How to calibrate a wheel sensor to within 1% #
Step 1: measure the rollout #
Chart values are a starting point, not an answer; tyre brand, width, casing and pressure all change the real rolling circumference. Common chart figures are about 2,096 mm for 700×23c, 2,105 mm for 700×25c, 2,136 mm for 700×28c and 2,155 mm for 700×32c.
- Pump the tyres to your normal riding pressure.
- Turn the front wheel until the valve is at the bottom, and mark the floor under it with tape.
- Sit on the bike, or have someone sit on it, so the tyre compresses as it does on a ride.
- Roll forward in a straight line for 5 or 10 full turns, until the valve is back at the bottom.
- Mark the floor and measure between the marks in millimetres.
- Divide by the number of turns. 10 turns over 21,050 mm gives 2,105 mm. Enter that as the wheel size.
Many GPS bike computers can also set the wheel size automatically from GPS over the first few kilometres, which gets you close. A careful rollout usually gets you closer.
Step 2: check it against GPS #
- Pick a long, open route, 20 km or more, with little tree cover, no tunnels and few tight bends.
- Record it with both the sensor and GPS at one fix per second, starting at the same spot.
- Compare the totals. Within 1%? You’re done.
- If there’s a steady gap: new circumference = old circumference × GPS distance ÷ sensor distance. If the sensor shows 40.8 km, GPS shows 40.0 km and your setting is 2,105 mm, the new setting is 2,105 × 40.0 ÷ 40.8, about 2,064 mm.
- Ride it again to confirm. A one-off gap can be a bad GPS day.
Don’t do this check on a forested or downtown route. Those are the places where GPS is the less reliable of the two, so “correcting” the sensor there makes it worse. Redo the rollout after any tyre change.
Does the same apply to a car’s trip meter? #
The principle is the same: a car’s odometer counts wheel turns, and a slope adds the same tiny amount. On a 6% grade at 100 km/h along the road, your horizontal speed is 99.8 km/h, a gap no display will show. If your car reads several km/h more than your phone on a mountain road, the reason is the speedometer’s built-in margin, which shows up everywhere; see why your car speedometer reads higher than a speedometer app.
Logging rides with a phone #
If you only need GPS, a phone on the bars logs a ride. In Speedometer GPS, press Start as you roll out, Stop to pause for a café break, and Save at home. The saved ride keeps its distance, top speed, average speed while moving, total and moving time, route and speed chart, and you can export it as GPX. The Trips screen adds saved rides into a lifetime odometer.
Know its distance limit before you compare it with a bike computer. To keep a parked phone from adding distance, the app ignores steps of less than 5 metres between one-second fixes, and fixes worse than ±20 metres. Above about 18 km/h (11 mph) that doesn’t matter. Below it, on slow climbs, in traffic and on easy spins, the trip distance reads short, so on a hilly ride it will usually trail a calibrated sensor for reasons that have nothing to do with the hill. Speed and average speed aren’t affected, and average speed × moving time gives a fair estimate. The app uses only the phone’s GPS and doesn’t connect to wheel or cadence sensors. For choosing between a phone and a dedicated unit, see phone vs bike computer.
Frequently asked questions #
Why does my GPS app show a different distance than my bike computer? #
The bike computer multiplies wheel turns by a circumference setting, and GPS adds up the distance between fixes. A wrong circumference, GPS cutting corners on twisty routes, and GPS dropouts or filtering under trees and on slow climbs account for nearly all of it. On open roads with a measured circumference, they usually agree within 1 to 2%.
Is GPS or a wheel sensor more accurate for distance? #
A sensor with a measured circumference, usually within about 0.5%, and it isn’t affected by trees, buildings or tunnels. GPS is typically within a few percent on open roads and needs no setup, which makes it a good check on the sensor.
Do hills add distance on a bike computer? #
Only a little: about 0.5% on a 10% grade and 2% on a 20% wall, and only on those sections. Over a whole hilly ride it’s usually a small fraction of a percent.
How do I measure my bike wheel circumference? #
Inflate the tyre to riding pressure, mark the tyre and the floor at the valve, and roll the bike forward several full turns with your weight on it. Measure between the marks in millimetres and divide by the number of turns.
Do I need to recalibrate after changing tyres? #
Yes. A different width, brand or casing changes the rolling circumference, often by 1% or more. Do a new rollout, or let a GPS computer’s automatic wheel-size setting relearn it.