skin temp: no approx raw value?, just deviation from assessed ‘baseline’?

[Victamon]

only relativized to “up to 30 nights of data” (shown in ⅒℉ ‑increment precision.. or I am guessing tenths of ℃ if you opt for metric units which would be less precise, unless that switches it to hundredths which would be much more precise), that presumably don't ever get overwritten with more recent data?

[Guy_]

Hi @Victamon - it's in tenths and rolling calculation of the baseline from the last 30 days.

Precision isn't that important because it depends more on how long your wrist is or isn't exposed to ambient temperature over night. 

You can get wide fluctuations but a trend is more significant and meaningful and is worth following up.