Readiness

What should today be?

A probabilistic read of your current physiological state. Every number here carries an interval — if the model is uncertain, it says so rather than rounding to false confidence.

Periodisation

Plan

Work back from your event into phases, each with its own targets per energy system. Prescribe reads this so it knows where in the block you are and how far out the day that matters is.

Strengths & weaknesses

Rider profile

What six months of your own riding says you are good and bad at — and what Prescribe is doing about it.

Is it working?

Progress

Three questions, answered from your own history: are you getting fitter, are you recovering, and are you riding enough.

State variables

Physiological state

Fitness and fatigue tracked separately per energy system, plus durability as its own slow variable measured from work-indexed power rather than assumed from FTP.

Conventional view

Training load

The familiar CTL / ATL / TSB picture, weekly work and hours, and your power–duration curve. Kept alongside Echelon's own state model so you can see both readings of the same history.

Adaptation markers

Recovery

Training load says what was demanded. These say what it cost and whether it is being absorbed. Every marker is scored against your own trailing baseline, and the readiness interval widens wherever they are missing.

Optimisation

Prescribe

Every candidate session scored on expected adaptation against expected fatigue and its downstream cost. The best session is rarely the hardest one.

Claude

AI coach

Sends the briefing below — your derived numbers only, never raw files — to Claude for a coach-style read. The model is told to respect the uncertainty intervals rather than talk past them.

Ingested activities

Rides

One row per activity, with the fingerprint Echelon retains: interval structure, energy-system load, work-indexed bests and decoupling. Raw streams are never stored.

Ingestion

Import

Drop a Strava bulk export straight in — Echelon reads the zip and the gzipped .fit files inside it without unpacking anything by hand. Everything is parsed in your browser; nothing is uploaded.

Reference

Help & glossary

Every abbreviation this app shows: what computes it, the scale it lives on, and how to read it. If a number here ever looks confident, this page is where you check how much it actually knows.

Baselines

Athlete

FTP drives the intensity normalisation behind every fingerprint. Changing it does not retroactively rewrite already-ingested rides — re-import to rebuild them.