How we measure
Aero tests are run under supervision at an indoor velodrome. This page says how a result is produced, what limits it, and what we decline to claim.
The method
Each step keeps a real change apart from everything else that moves between runs.
- 01
Supervised velodrome runs
Every run is ridden on an indoor velodrome under supervision. Still air and a fixed lap length remove most of what changes outdoors.
- 02
Lap-triggered straights
A trigger at the line marks each lap. The analysis reads the straights between triggers, so the bends are kept out of the number.
- 03
Energy balance per interval
For each short interval the analysis balances the energy the rider puts in against the energy lost to air, to rolling resistance and to changes in speed. The drag follows from that balance.
- 04
Baselines at both ends
The baseline setup is ridden at the start of the session and again at the end, so slow drift in the conditions or in the rider is bracketed instead of being mistaken for a change.
- 05
A confirmation run
A setup that looks faster is ridden again, with a baseline run beside it, before it is reported as a gain. A good first run is a lead. A repeat that agrees is the result.
What adds uncertainty
These are the places a result can be off. Each is controlled where it can be, and the rest is carried into the range. The size of each is not quoted here.
- Power meter
- Measured power is the main input. A meter that reads a little high or low shifts every run the same way, so it matters most when results are compared across days or across meters.
- Speed and wheel
- Speed comes from the wheel. A small error in the wheel's circumference scales every speed and distance in the run.
- Air density on site
- Drag depends on the density of the air, which follows temperature, pressure and humidity. We take readings at the track on the day. Without them, runs can still be compared with each other, but the result is not given as an absolute number.
- Wind indoors
- Indoors the air is nearly still, but doors, ventilation and other riders can move it. The test day records anything that could.
- Rider position repeatability
- No rider holds the same position twice. This lap to lap spread is read from the laps themselves, and it is the largest part of the range on a result.
What we refuse to claim
A report says less when the evidence is thin, and it says why.
- A result without its range.
- A change called while it sits inside the noise.
- A figure from outdoor testing. Aero results come from indoor velodromes only.
- A watts figure from a run that was not confirmed.
- A difference between runs that had no lap triggers. Without triggers the run cannot be placed on the lap, so it does not take part in a comparison.
- An absolute figure from a run without air readings taken at the track. Such a run is compared only inside its own session.
The smallest change we can detect
The smallest change we can detect will be published after the first calibration day at the track. Until then no such figure is quoted here or anywhere else on this site.
Tested against known answers
The method has been tested on simulated rides where the true answer was set in advance, including rides with sensor lag, uneven pacing and missing data. The sample report is one of them.