I think that we have speculated before that data from Sam's watch showed her heart rate/blood pressure/etc suddenly spike (during a sudden attack) and then all of them slowly ebb away (as she died).
IIRC a couple of other cases where this kind of evidence was shown were given as reference points, back in the thread.
I gave one South Australian case as a reference ... will try to find the link again and add it.
ETA link:
Ms Matteo said the data showed activity consistent with Myrna Nilsson being ambushed and attacked as she walked into her home just after 6:30pm.
She said it then showed activity and heart rate measurements consistent with her body going into shock and losing consciousness.
"The deceased must have been attacked at around 6:38pm and had certainly died by 6:45pm," she said.
Smartwatch data helped police make arrest in Adelaide murder case, court hears
I think we have to be careful with assumptions.
For instance with my garmin watch and iphone, some of the run data is syncing in realtime to my phone - particularly steps
However the activity data does not sync until i end the activity. So I can imagine a situation where the phone was switched off, and activity not ended - you might not get all the data synced to the phone and if you don't have the watch, then you never get that data.
tldr; it can be complicated