Cadence Labs

Turn performance data into clear execution.

No black boxes. No secret formulas.

Understand the exact principles behind every watt, split and fuelling target.

Inspect the platformRequest team walkthrough

Most performance tools hand you a single number and ask for blind trust. Cadence Labs shows the working: the source behind every constant, the assumptions behind every output, and the confidence interval around every target.

Total transparency

Inspect the source behind every constant, and see the true uncertainty behind every projection.

Configurable to your protocols

Adaptable to your organisation's own physiological datasets and testing standards.

Unified forward and inverse modelling

The same physical and bioenergetic model that designs the pacing strategy also reconstructs the ride file afterward, so planning and post-ride validation speak the same language.


Built for the whole performance staff: coaches, nutritionists, sports scientists and performance directors, all working from the same numbers.


What each part of your staff gets

One model. Four different views into it, each built for how that role actually works.

Nutritionists

Carbohydrate and fat oxidation demand, worked out sample by sample through the ride rather than at the average, and reported as a rate in grams per hour, the number a fuelling plan is actually built from.

Sports scientists and coaches

Aerodynamic drag, rolling resistance, gravity and drivetrain loss modelled together, and a rider's own drag position solved from their ride files rather than assumed from a photo on the bike.

Coaches and DSs

How closely a session matched what was prescribed, interval by interval, next to the metabolic cost of getting there. Plan against actual, on the same terms.


One model, not three separate tools

A pacing app, a telemetry platform and a fuelling spreadsheet each carry their own idea of what a ride costs. When they disagree, nobody can say which one is right. Cadence Labs runs one model, both ways.

Before the ride, the model works out what a session will ask of a rider, the power, the course, the fuel, from the same arithmetic every time.

After the ride, that same model takes the file that comes back and rebuilds what actually happened, sector by sector.

Because it is one model both ways, what was planned and what happened are measured on identical terms. A gap between them is something that happened on the bike, not a disagreement between two tools that were never working from the same numbers.

How the engine works

Full visibility into every number

A figure a coach cannot check is a figure nobody can stand behind in a team meeting.

Nothing is filled in

If a channel is missing from a ride file, heart rate, power, whatever it is, it stays missing. The engine never invents a value to close the gap.

Every figure is labelled

Measured on this ride, tested on this athlete in a lab, or a population figure used because nothing more specific exists yet. The three are never shown as if they were the same thing.

Confidence is stated, not implied

Every output carries the range it could reasonably fall within, so a number is never presented as more certain than the data behind it allows.


What it will not do

These are not gaps. They are lines the engine holds on purpose.

No food, no products

Output stops at a demand, expressed as a rate in grams per hour. What a rider actually eats or drinks is a nutritionist's call, never ours.

No coaching prescription

The engine reports what a session demanded and what it cost. It does not tell a coach what to do about it. That decision stays with your staff.

No labelling a rider from population data

A figure drawn from published research describes a population. It is never turned into a statement about one named athlete. That judgement belongs to whoever knows the rider.


Two ways to use it

The same engine, delivered two ways. Which one depends on whether you are a coach or an organisation.

Session View

For cycling coaches working with their own riders.

  • Analysis
  • Reports
  • White-label

Embedded

For teams, academies, federations and platform partners.

  • Single-tenant
  • Configurable
  • API

How we handle sources

Every constant the engine uses is registered against the source it came from, and no part of the engine can promote a value above the evidence behind it. Where we have not yet checked a constant against the paper it came from, the report says so on its face rather than letting it pass as established.

Methodology