What an average speed does and does not tell you
Two riders both report 15.5 mph. One rode solo into a crosswind on a rolling course; the other sat in the third wheel of a group on a flat loop, where drafting cuts aerodynamic drag by thirty to forty percent. The power outputs are not close. Add the recording question on top: an app averaging over elapsed time including traffic lights reads two to three miles an hour below one averaging over moving time, on the same ride.
The band shown in the results assumes solo riding on flat ground with moving-time averaging, and it is offered as orientation rather than a grade. The only comparison that carries information is your own times on the same route in similar conditions.
Gear inches, and why the unit survives
Speed is ratio times wheel circumference times cadence. Gear inches expresses the first two as a single figure: the ratio multiplied by the wheel diameter, giving the diameter of the direct-drive wheel that would move you the same distance per pedal turn. It is an anachronism from the penny-farthing era and it persists because it is genuinely useful — one number that lets you compare a road bike's 50/17 against a touring bike's 44/15 against a mountain bike's 32/11 without holding three wheel sizes in your head.
On 700c, 50/17 is about 77 gear inches and rolls 6.2 metres per pedal revolution, which at 90 rpm is 20.7 mph. Development in metres is the same information for people who prefer thinking in distance covered rather than a virtual wheel.
| Gear | Gear inches | Speed at 90 rpm | Where it belongs |
|---|---|---|---|
| 50 / 11 | 119 | 32 mph | Descending, sprinting |
| 50 / 17 | 77 | 20.7 mph | Flat cruising |
| 34 / 17 | 53 | 14.1 mph | Rolling terrain |
| 34 / 28 | 32 | 8.5 mph | Sustained climbing |
| 34 / 34 | 26 | 7.0 mph | Steep, loaded, or tired |
Cadence is a gear-selection problem
The common beginner pattern is a heavy gear turned slowly, often below 60 rpm. It feels productive and it loads the knee substantially — pedal force at 60 rpm is half again what it is at 90 for the same power — while recruiting fast-twitch fibres that fatigue quickly. Most riders settle between 85 and 95 rpm on the flat, and the specific number matters far less than the habit of shifting to hold it as the terrain changes.
On a climb, cadence dropping below about 70 rpm is the signal that you have run out of gears rather than a sign to push harder. If 34/28 is still forcing you down to 60, a larger sprocket is the answer, and modern rear derailleurs will generally take a 32 or 34 without complaint. There is nothing to prove in a gear that is too big; the people who ride fastest uphill are on the smallest gears they can justify.
What actually makes you faster, in order
Above about 15 mph, aerodynamic drag is the dominant resistance and it rises with the square of speed, so the ordering of improvements is not intuitive. Position comes first and costs nothing: bending the elbows and dropping the torso reduces frontal area more than any component will. Tyres come second — supple casings and correct pressure can be worth tens of watts against cheap tyres, which is more than most wheel upgrades. Consistent riding, three times a week, comes third. Frames and wheels come after all of that.
Climbing is a separate problem governed by power divided by total mass, where aerodynamics barely feature. That is the arithmetic behind the observation that losing weight from the rider is the cheapest available upgrade, and also behind the fact that it is a poor idea to pursue very far.
Related
Setting effort by heart rate rather than speed sidesteps most of what makes average speed unreliable: heart rate zones. For a comparison of energy cost across activities on the same MET basis, see walking and stair climbing.
Questions people ask
How do I measure cadence without a sensor?
Count how many times one knee reaches the top of the stroke in fifteen seconds and multiply by four. Do it on a flat section at a steady effort, since the number drifts as soon as the gradient changes. A cadence sensor is inexpensive and most bike computers and several sports watches will pair with one, which is worth it if you intend to work on cadence rather than just check it occasionally.
The gear numbers assume 700c. What about other wheels?
Scale by rolling circumference. The calculator uses 2,096 mm for a 700 x 25c. A 700 x 28c is about 2,136 mm, so speeds run roughly two percent higher than shown. A 26 inch mountain bike wheel is near 2,050 mm, about two percent lower. A 20 inch folding bike wheel is around 1,490 mm, which makes every speed on this page roughly thirty percent optimistic — the same gear ratio on a small wheel is a much lower gear, which is why folders run enormous chainrings.
Does commuting by bike count as exercise?
A commute of thirty minutes each way covers the standard weekly moderate-intensity aerobic guidance on its own, and cycle commuters show consistently lower cardiovascular risk across large cohort studies. The qualifier is intensity: a relaxed nine mile an hour ride on a flat path sits near 4 METs, which is at the bottom edge of moderate. If the commute is meant to be the training, choosing a segment to ride hard turns it into an interval session without adding any time.
Why is my computer calorie figure so different from this one?
Because it is probably using a different method. A speed-banded MET estimate, as used here, knows nothing about wind, gradient or whether you were drafting, and cycling speed is an unusually bad proxy for effort for exactly those reasons. Heart-rate-based estimates track your physiology but drift with heat, caffeine and fatigue. A power meter measures mechanical work directly, and kilojoules convert to kilocalories almost one to one because gross cycling efficiency is close to 24 percent. If the number matters to you, that is the instrument that answers it.