Three things that have to be true

1. 1

   Does the watch see anything real?

   reinforced, with one honest nuance

   The watch is producing signal, not expensive noise.

   A consumer watch on one wrist could, in principle, be producing numbers that mean nothing. The strongest answer available right now is the shape of what it records on a bad day: when a crash hits, several channels move together (resting heart rate up, body battery's floor down, stress up) graded by how bad the day is, in the directions physiology would predict. Noise doesn't organise itself into a coherent, severity-scaled fingerprint like that.

   One honest caveat, and it cuts two ways: these channels are not independent witnesses; they share an autonomic cause, and collapse to a single factor measured several ways. For counting tests, that's a discount, not a chorus of seven. But a single factor that moves coherently and in proportion to severity on bad days is also the shape a real signal takes, noise has no reason to fold up that neatly. So the weight comes from the coherence and the grading, not from counting channels. What this step can't do on its own is show the factor tracks anything outside the watch, and that check has now come back. The two anchors it was waiting on both landed: the watch's factor moved significantly during the one independently-known autonomic event in the record (a 14-day COVID infection), and the wider consumer-wearable literature was reviewed. The COVID check is a qualified pass (carried by the stress and body-battery channels, with the heart-rate channel masked by peak fitness at the time) so the link is now reinforced by evidence from outside the watch, not just its own internal coherence. Still short of 'solid', and honest about the nuance, but no longer the link leaned on most lightly.

   seven signals, or one? → · the COVID check →

2. 2

   Does what it sees track how I actually feel?

   established

   The watch's signals and the felt-state score pick up the same days, in proportion to severity.

   A crash defined purely on the felt-state score also shows up on the watch: roughly a standard deviation above my personal baseline on resting heart rate and on overall stress, about one and a half below baseline on body battery's daily low. Dip days move the same channels the same way, about a third as far. Ordinary days sit near baseline. The two measurements pick up the same days, scaled by how bad they were, and that overlap is what makes it legitimate to analyse them together.

   But not the watch validating the score. They're two measurements of one body that share an upstream cause, so when they agree, they're agreeing about the same physiology through different machinery, not confirming each other. And the relationship isn't a straight line: line the numbers up day-by-day and it nearly vanishes; line them up around crashes and it's clearly there.

   not a straight line →

3. 3

   Do Wiggers' specific patterns hold in my body?

   in progress

   Each channel she names, tested against four years of my data, pre-registered, one at a time.

   This is the actual question the site asks, and the only one of the three that's still open, test by test. Some of her patterns show up. Some don't. Most are still being judged on the whole record. That open question, tested pattern by pattern, is the living part of the site, and the two links above are what earn it the right to exist.

   the field guide →

This chain is a scaffold, not a proof. It shows where the argument stands (one link solid, one reinforced now by a check from outside the watch, one still open) so you can see exactly how much weight each part of the site is allowed to bear.