endurance-domain-referencelisted

# Endurance domain reference **Use this when** you hit an endurance term or metric (CTL, EF, decoupling, CSS, TID, taper, dose-cycle, carb g/h…) and need its plain meaning, where the repo computes it, the normal ranges this repo uses, and the honest caveats — or when you are reading/writing `insights/` or `coach/` code and need the physiology. **Don't use this when** you need the *statistical machinery* that proves an n=1 signal is real (effective-N, Fisher-z CIs, Benjamini–Hochberg FDR, walk-forward, permutation nulls) — that lives in `endurance-coach-proof-and-analysis-toolkit`; or the *code contracts* (Provenanced fields, the write gate, data flow) — that lives in `endurance-coach-architecture-contract`; or the *go/no-go campaign* to validate a detector — `endurance-coach-n1-validation-campaign`. --- ## Two domain doctrines that override everything below These are not opinions. They are hard rules the codebase and coach persona are built on. > **1. Priors, not laws — this athlete's n=1 data outranks the textbook.** Everything in > `knowledge/sports-science.md` is a *population* prior (small samples, modest effects, big individual > variation). Individual training response is roughly 50% heritable with 20–45% non-responders, so no > a-priori dose-response model is trustworthy for one person. Where a prior disagrees with THIS athlete's > own pulled data, **the prior yields.** See `knowledge/sports-science.md` header. > **2. Defer to AI Endurance's model — never re-der