Bike Odometers: How Your Mileage Is Measured and Why Totals Disagree
You and a friend roll out together, ride the same loop, stop at the same café — and the two computers on your bars disagree about how far you just went. Neither of you took a detour. It happens often enough that Strava keeps a help page devoted to the question. A bike odometer measures distance one of two ways — counting wheel revolutions and multiplying by a wheel-circumference number, or joining up GPS position fixes and summing the gaps — and because those two methods go wrong in different directions, two computers on the same ride rarely agree. Here is what each one is doing, which errors are yours to fix, and how to keep a lifetime total when you change devices.
Key takeaways
- A wheel-sensor odometer counts revolutions and multiplies by wheel circumference — one number you enter decides every ride you ever record.
- Garmin's wheel-size table lists 700 × 23C at 2096 mm and 700 × 25C at 2105 mm. Pick the wrong line and every distance is off by about 0.4% — roughly 430 m per 100 km, always the same direction.
- GPS distance is built from position fixes and, per Strava's documentation, assumes a flat surface: vertical movement from topography is not counted. A wheel sensor does pick it up.
- Drift and blocked or reflected signal make rides read long or short. GPS.gov puts a smartphone near a 4.9 m (16 ft) radius under open sky, worse by buildings, bridges and trees.
- Honest limit on our own kit: the Aoocci NAV 1 (around $140, as of July 2026) has no GPS receiver of its own — position comes from your phone via the companion app — and its listing includes no wheel-sensor input, ANT+, power, heart rate or cadence. Navigation is also a subscription: the listing gives a 3-month free trial, then pricing from around $3/month with quarterly and yearly options, so around $140 is not the whole cost of running it.
What a bike odometer is actually counting
The odometer on a bike computer is a running sum of every ride it has recorded. The question worth asking is where each ride's distance came from, and there are only two answers: something watching your wheel turn, or something watching satellites. Strava, which ingests rides from a wide range of devices, describes the first path plainly — ground-speed distance comes from counting wheel revolutions and multiplying by the wheel circumference. The second calculates distance from GPS data as you ride, then re-parses it after upload. Both look equally confident on a handlebar screen; only one is measuring your actual wheel.
The wheel method: one number decides everything
The classic version is a magnet zip-tied to a spoke and a sensor on the fork leg — each pass of the magnet is one revolution. Newer sensors drop the magnet: Garmin's Speed Sensor 2 attaches to the hub of either wheel, self-calibrates when paired with a Garmin Edge, and keeps capturing data even when you ride without a head unit, on a claimed battery life of up to a year.
Whatever does the counting, the arithmetic hangs on one input: circumference. Garmin's own wheel-size reference lists 700 × 23C at 2096 mm and 700 × 25C at 2105 mm — nine millimetres apart. Set 23C while you are actually riding 25C and every distance runs short by about 0.4%, roughly 430 m per 100 km, forever, in the same direction. That is the character of wheel-sensor error: not random noise, a fixed bias baked into the whole log.
The fix is a roll-out. Sheldon Brown's calibration guide explains why chart values only get you so far: numbers from a chart or derived from ISO/ETRTO sizes are generally accurate to within one or two percent, and tread thickness, tyre pressure and rider weight all change the effective size — so rolling circumference should be measured with the rider aboard. The guide puts actual rolling diameter about 1% smaller for a road tyre, smaller still at low pressure. Mark the valve, roll several revolutions against a metal tape, divide, and the number belongs to your bike, not to a table.
The GPS method: joining the dots
The GPS path skips all of that — nothing to calibrate, nothing to zip-tie. The device logs a stream of positions and adds up the distance between them. That convenience costs two things.
The first is geometry. Strava's documentation is explicit that its GPS distance assumes a flat surface and does not account for vertical speed from topography, while a wheel sensor captures the extra distance elevation adds. The gap is modest — a sustained 10% grade adds only about 0.5% to the ground you cover versus its map length — but on a climbing day it pushes the GPS number under the wheel number, not over it.
The second is position error, and that one is not modest. GPS.gov puts GPS-enabled smartphones at typically within a 4.9 m (16 ft) radius under open sky, worsening near buildings, bridges and trees. Strava's help pages describe needing roughly seven or eight satellites for about 10 m of accuracy, and note that buildings, trees, tunnels and mountains block signal while reflections off buildings confuse the receiver. The upshot: drift, signal loss or a jumpy track can make an activity report more or less distance than you actually travelled. Long or short — GPS error does not politely pick one direction.
Our own NAV 1 belongs to this family, one step further out. Its positioning source is your smartphone's GPS through the companion app; there is no receiver inside the unit. It records speed, mileage, altitude and trip duration for review on the phone. Four honest consequences follow. Leave the phone behind, run it flat or lose the app connection and there is nothing to plot or count. Its 600 mAh battery is rated at about 6–8 hours of active use, so a century or a brevet day can outlast it. And because the listing includes no wheel-sensor input, ANT+, power meter, heart rate or cadence, you cannot hand it a rolled-out circumference — it is a glanceable ride computer, not a training head unit. Finally, the navigation service is a subscription: the listing gives a 3-month free trial, then pricing from around $3/month with quarterly and yearly options, so the around $140 is not the whole cost of running it.
Why the two totals never match
| What is happening | Wheel-sensor odometer | GPS-derived odometer (including phone-fed units like the NAV 1) |
|---|---|---|
| How distance is produced | Counts wheel revolutions, multiplies by wheel circumference | Joins successive position fixes and sums the gaps |
| What you have to get right | The circumference — from a chart, a roll-out, or auto-calibration | Nothing to set; you are at the mercy of sky view |
| Climbing | Captures the extra distance elevation adds | Assumes a flat surface; vertical from topography is not counted |
| Trees, tunnels, city streets | Unaffected — the count is mechanical | Blocked and reflected signal cause drift; rides read long or short |
| Character of the error | A fixed bias, same direction every ride (chart values ~1–2%) | Varies with conditions ride to ride (phones ≈4.9 m radius, open sky) |
| Needs a phone or satellites? | No | Yes — the NAV 1 specifically uses your phone's GPS via the companion app |
Read the table as two instruments rather than a right answer and a wrong one. If your friend runs a hub sensor and you are on phone-fed GPS, a hilly ride should come out differently — theirs counts the extra ground the slope adds, yours measures the flattened map version. Add a 1% circumference error on their side and a tree-lined valley on yours, and the gap stops being mysterious.
Which number should you trust?
For almost everything riders do with mileage — chain and tyre intervals, this month against last month, a lifetime figure ticking up — consistency matters more than absolute truth. One method, one set of settings, all year: a log that is quietly 1% short every time is more useful than one that is dead accurate on Tuesday and drifting under the trees on Sunday. When the absolute number does matter — a fixed-distance route, a course you are measuring for someone else — the wheel sensor is the more defensible instrument, because its error is one measurable bias you can remove with a tape measure. GPS error is environmental and changes with the ride, so there is nothing to correct.
What the rider feedback says
Most of the rider feedback we track comes from motorcyclists, and there the complaint about bar-mounted phones is not accuracy — it is vibration eventually killing the phone's camera, even through a damped mount. One rider writing about a handlebar phone mount put it this way: "I don't want to kill another around $1,200 phone." A bicycle is gentler on hardware, but the logic carries: better a sealed 52.7 g unit takes the road buzz — even when, as with the NAV 1, the phone in your jersey pocket is still supplying the GPS.
Keeping your lifetime total when you change computers
Here is what people discover too late: the lifetime total usually lives on the device, not in the cloud. Retire the unit, factory-reset it or drop it in a ditch and the number goes with it. Platforms do not automatically rescue you either — Strava's gear tracking only accumulates mileage that was recorded on Strava, you cannot type in a starting figure, and the documented workaround is a manual activity assigned to that bike to stand in for everything that came before.
So: photograph the odometer screen before any reset, firmware wipe or handover. Check before buying whether a head unit lets you enter or edit a starting total — do not assume it can. And decide early who owns the lifetime number — the device, an app, or a spreadsheet — because that decision is far cheaper to make at 200 km than at 20,000. On the NAV 1, the listing puts ride data including mileage in the companion app for review, so the phone side is where the history sits. The recurring fee is described on that listing as a navigation subscription; whether anything about the logging changes once the 3-month trial ends is not something the listing settles.
For the mechanics underneath all this, our explainer on how GPS actually works covers what a fix really is, and why GPS speed and your speedometer disagree is the speed-side version of the same argument.
Aoocci NAV 1 Smart Cycle Computer — around $140 (as of July 2026)
A 1.43-inch round IPS touchscreen at 426 × 426 and 1000 nits, 52.7 g and 13 mm thick, IP67-rated by the manufacturer, logging speed, mileage, altitude and trip duration for review in the companion app. It ships with two mounts — an ACC direct-power mount for a motorcycle and a standard bicycle mount — with a quarter-turn Twist-Lock to move it between them. Positioning comes from your phone's GPS through the app: there is no GPS receiver in the unit, and no wheel-sensor input. Navigation runs on a subscription — a 3-month free trial, then from around $3/month, with quarterly and yearly options — so budget for that on top of the around $140.
See the NAV 1 →Running it on two bicycles rather than a bike and a motorcycle? A spare NAV1 Mounting Base with 25 mm ball head is around $15 (as of July 2026) — but read its fitment notice before you count on it. It is an accessory-only listing: it includes the mounting base and two nuts, while the NAV 1 device, the handlebar clamp and the mounting arm are all excluded. To put the unit on a second bicycle you would still need a clamp and a mounting arm whose socket takes a 25 mm ball head, so around $15 is the spare base, not a working second mount.
Frequently asked questions
How does a bike odometer measure distance?
One of two ways. A wheel-sensor odometer counts wheel revolutions — traditionally with a magnet on a spoke, now often with a magnetless hub sensor — and multiplies that count by the wheel circumference set on the device. A GPS odometer instead logs a stream of position fixes and adds up the distance between them. The lifetime figure is just the running sum of every ride recorded by whichever method your computer uses.
Why do my bike computer and my friend's show different distances for the same ride?
Because they are measuring different things. If one uses a wheel sensor and the other GPS, a hilly ride differs by design: Strava's documentation notes that GPS distance assumes a flat surface and does not account for vertical movement from topography, while a wheel sensor captures the extra distance elevation adds. On top of that, a wheel sensor with a slightly wrong circumference is biased the same way on every ride, and GPS drift, blocked signal or a jumpy track can make a ride read long or short.
Is GPS mileage or wheel-sensor mileage more accurate?
Neither is automatically right, but their errors behave differently. A wheel sensor's error is a fixed bias you can remove: roll the wheel out with your weight on the bike and enter the real circumference. GPS error depends on where you ride — GPS.gov puts smartphones near a 4.9 m radius under open sky and worse by buildings, bridges and trees — so it varies from ride to ride with nothing for you to correct. For maintenance intervals and comparing your own rides, consistency matters more.
Does the Aoocci NAV 1 have its own GPS?
No. The NAV 1's positioning source is your smartphone's GPS, delivered through the companion app, so both its distance and its navigation depend on the phone being with you, charged and connected. It records speed, mileage, altitude and trip duration for review in the app. Its listing includes no wheel-sensor input, ANT+, power meter, heart rate or cadence, so you cannot calibrate its distance with a rolled-out wheel circumference.
How do I keep my lifetime mileage when I switch bike computers?
Assume nothing transfers. The total usually lives on the device, so photograph the odometer screen before any reset or handover, and check before buying whether the new unit lets you enter a starting total — do not assume it can. Platforms are not an automatic safety net either: Strava's gear tracking only counts mileage recorded on Strava, you cannot manually set a gear's mileage, and the documented workaround is a manual activity assigned to that bike.
