Evinlo Explainers
How Evinlo works
Every number in Evinlo comes from rules you can read. Each explainer takes one area from what your Apple Watch records to what you see in the app, in a few minutes of reading.
Each one starts with the one-minute version, then goes through the steps with a figure for each, a worked example, what you see in the app and the honest limits. The last pages are for experts: every parameter, formula and source. The figures and examples are calculated by Evinlo’s own algorithms on made-up people, never on anyone’s real data.
-
Evinlo, explained
The whole system: from Apple Health through the overnight layer to each area and Home, with your own usual ranges, privacy on your iPhone, versions and the evaluation tool, and a map of the other explainers.
The question it answersHow does Evinlo turn what my Apple Watch records into what I see in the app?
Covers every area
Read the PDF: Evinlo, explained, 15 pages, 1.4 MB -
Sleep, explained
The sleep score, your sleep need and sleep debt, tonight’s need, naps, and what your answer to “How did you wake up?” changes.
The question it answersWhy did last night get this score, and where does tonight’s need come from?
Algorithm version 1.3.0
Read the PDF: Sleep, explained, 15 pages, 1.5 MB -
Alertness and the daily check-in, explained
A published model of alertness, driven by your sleep times and tuned to your own one-tap answers.
The question it answersHow awake may I feel today, and why does Evinlo ask me first?
Algorithm version 1.3.0
Read the PDF: Alertness and the daily check-in, explained, 11 pages, 935 KB -
The overnight layer, explained
Which nights count, how each night’s values are taken from Apple Health, how your usual ranges form, and the heart signal.
The question it answersWhat does “your usual” mean, and which nights does it come from?
Algorithm version 1.0.0
Read the PDF: The overnight layer, explained, 14 pages, 1.5 MB -
Body signals, explained
Overnight HRV, sleeping heart rate, respiratory rate, wrist temperature and blood oxygen, each against your own recent nights.
The question it answersWas anything my body recorded last night outside my usual range?
Algorithm version 1.0.0
Read the PDF: Body signals, explained, 11 pages, 1.4 MB -
Overnight changes, explained
A narrow check for several overnight signals moving away from your usual together. It shows that something changed, never why. Not a diagnosis.
The question it answersDid several of my overnight signals change together?
Algorithm version 1.0.0
Read the PDF: Overnight changes, explained, 13 pages, 1.5 MB -
Readiness, explained
A score from 0 to 100 for how ready your body is for effort today, from your night, your overnight signals and yesterday’s activity, against your own usual.
The question it answersHow ready is my body for effort today, and what moved the score?
Algorithm version 1.0.1
Read the PDF: Readiness, explained, 12 pages, 995 KB -
Stress, explained
Recovery and stress across the waking day: when your heart rate at rest was above your own usual level, and for how long. There is no daily stress number.
The question it answersWhen was my heart rate at rest above my usual level today?
Algorithm version 1.0.0
Read the PDF: Stress, explained, 14 pages, 1.6 MB -
Movement, explained
Steps, active energy, exercise minutes and time in daylight against your own usual, with no score, no goal and no ring.
The question it answersHow does today’s movement compare with what is usual for me?
Algorithm version 1.0.0
Read the PDF: Movement, explained, 14 pages, 1.3 MB -
Fitness trends, explained
Estimated VO2 max, cardio recovery, walking heart rate and resting heart rate over six months, as directions, not verdicts.
The question it answersIs my fitness changing over the months, or is it just noise?
Algorithm version 1.0.1
Read the PDF: Fitness trends, explained, 12 pages, 1.3 MB
The complete reference
The explainers follow docs/scoring.md, the scoring document next to the code on GitHub. It describes every area in full: each rule and parameter, the data it reads, the studies it rests on and what is a design choice, how each version was validated, and the questions still open.
If an explainer ever says something different from the code, the code is what the app runs, and the difference is a mistake worth reporting. To question a rule or propose a better one, see how to contribute, or send feedback.