Poor Resources algorithm
-
Resources show 1% or even 0%, which means I should be dead.
Please improve your algorithm and assume no activity (and therefore rest) when the watch is turned off.
I only turn on the watch for sport activities in order to save battery and most of all because I don’t want nor need a device on my wrist when I sleep (it hinders my ability to sleep well…)
-
@galerna the watch usually starts with 70% after turned back on again, doesn’t it?
-
@galerna why do you even care about Resources if you only use watch to track workouts? It is clearly designed not for your use case.
-
@galerna do you wear it while sleeping? It starts from a certain value after turning the watch on (idk how that value is calculated) but its intended use is to wear it overnight where it charges during sleep and then while wearing it during the day it discharges when your active/stressed and sometimes charges again when you’re relaxing enough, e.g. taking a nap.
It’s not working correctly when you just use the watch during exercise -
@ChrisA said in Poor Resources algorithm:
It’s not working correctly when you just use the watch during exercise
That’s what I mean: please fix the algorithm so it works in all cases. It’s as simple as just assuming the user is resting (not exercising) when the watch is off.
-
@galerna
the watch can only make predictions based on what its measuring. if it has no data it can’t guess your resources -
@galerna said in Poor Resources algorithm:
That’s what I mean: please fix the algorithm so it works in all cases. It’s as simple as just assuming the user is resting (not exercising) when the watch is off.
I don’t believe the algorithm is as simple as either going down at one constant rate or increasing at one constant rate. Let’s consider a person who middle of the day at noon has Resources at 50% and pretend there’s “four paths” the algorithm can choose: decrease 1% per 10m, decrease 5% per 10m, increase 1% per 10m, increase 5% per ten minute. And let’s make this a multiverse where we can model each path.
At 13:00, the four options are:
- 50 - 6 = 44%
- 50 - 30 = 20%
- 50 + 6 = 56%
- 50 + 30 = 80%
This is very simplistic, of course, and doesn’t even present a model where resource just “stay still” and would still be at 50% rather than these four. There’s likely dozens of possible outcomes.
Just saying “the user is resting when watch is off” doesn’t give any level of detail. Is the user asleep the entire time? If so, is their breathing hard and labored due to oncoming illness or slow and steady? More likely, “resting,” might mean sitting on a couch. Is the user reading a book? Watching a horror movie and getting nervous?
We’re establishing the watch is off and taking no readings, so it has no idea of any state, so any info it has is unlikely to be correct. For people like you who only wear the watch for sessions, Resources cannot give you any believable data at all and should be ignored. It’s like asking it to record your outdoor run without being turned on, and expecting it to guess if you did 5km or 10km as well as pace and HR without having collected data. How is that supposed to work? It’s the same for resources — no inputs, no proper calculation.
