J Data · Method guide
J Metrics are JMM names for clearly described calculations. JStress uses the published session-RPE method. These names do not turn a training log into a laboratory test.
Completed minutes × session RPE
The reported internal load of one completed session, in arbitrary units (AU). Add completed sessions to describe a day.
Limit: Requires actual duration and your whole-session effort rating. It is not a power measurement or a performance score.
42-day exponentially weighted daily JStress
A slower-moving description of your recorded training load. Recent days receive more weight.
Limit: It does not measure aerobic fitness, strength or injury resistance. A short or incomplete history weakens interpretation.
7-day exponentially weighted daily JStress
A faster-moving description of recent recorded load on the same session-RPE scale.
Limit: It does not measure physiological fatigue, illness or recovery directly.
J Base − J Recent
A comparison between the slower and faster load trends. Negative means recent load is above the longer trend; positive means it is below.
Limit: Neither sign guarantees readiness. There is no universal safe, optimal or race-ready target.
The 7-day and 42-day settings are JMM model choices, not personal biological constants. Read them as load trends, alongside history completeness, symptoms, sport demands and performance tests.
A day is known only when every recorded completed session has actual minutes and a reported RPE, or when you explicitly confirm a rest day with no completed training. Confirmed rest contributes zero. An unrecorded day is unknown.
For each uninterrupted sequence of known days, the first daily total seeds each average. Later days use: new average = previous average + α × (daily total − previous average), where α = 1 − exp(−1/τ), with τ = 7 for J Recent and 42 for J Base.
An unknown day resets that sequence. J Recent appears after 7 consecutive known days; J Base and J Balance appear after 42. These are JMM display rules, not validated safety thresholds. Individual session JStress can be shown sooner.
Do not mark an unknown day as rest just to reveal a chart. Correct missing sessions and effort ratings so the trend reflects the work you actually completed.
Illustration only — not your training data
45 completed minutes × session RPE 4 = 180 AU JStress. A 30-minute session at RPE 6 also gives 180 AU, but the sessions can produce different adaptations and mechanical strain.
Ask the AI coach: “Which sessions and dates support this trend? What is missing? Why does this affect today’s session?” A useful answer explains the evidence and uncertainty. It cannot diagnose an injury or guarantee a race result from a load score.
Session-RPE is a published method, not a JMM invention. Foster and colleagues evaluated it during cycling and basketball. That supports using perceived effort to describe load; it does not validate a new JMM composite, every sport-specific prescription, or an injury forecast. Foster et al., 2001 — PubMed.
TrainingPeaks describes its own Training Stress Score methods. References to third-party metrics identify those methods and do not imply affiliation. JMM naming is not a statement of trademark registration or legal clearance.