The recent update included a feature to indicate that a device is not on. All we have been told about this is that “…if Sense is showing a device as ‘On’ but you know it’s not, you can give us a heads up”. Most frequently, this occurs when several devices show up as a single device, e.g. when a coffee pot and hair dryer both display as “unknown heat”. It does not seem helpful for me to notify you that the device is not on, when I know that it is probably some other device with a similar footprint. Should I reserve the feature to use when Sense tells me specifically that a certain device is on, e.g., a fridge, and I know it is some other device? Even in that case it seems more helpful if I can simply rename the device to what I know it to be. Can you give us a specific use case when it is appropriate to use this feature?
I have a light in the bathroom (halogen bulbs) that Sense always detects when it turns on. However, it does not always detect when it turns off. I plan to use this button for when situations like this occur.
Same here - happens with a light and the fridge. Problem is catching when it happens - hopefully a message will prompt them to review the data, because I usually only catch it hours after the fact.
Wonder if it is an issue with the halogen signature. I have a torch light with a 500 watt bulb and had several instances where it would be turned off yet Sense kept it in the on status for ten minutes of so after it went off.
Not sure, but here are a couple instances.
Fridge:
I think that’s actually two cycles, but it missed the end of the first one and carried over to the second cycle.
(CFL) lights in washroom:
I’m certain this one is an error since the light is on a 5s timer
IDK. Sometimes it works correctly. Sometimes it says that light has been on for hours when it has actually only been on for about 10 minutes.
How do we get to report this? I haven’t see any option to report the issue.
“report a problem” down at the bottom:
(note you can also see a couple instances of the “hangover effect” in that graph)
I think it would be more helpful to be able to target an instance in the past that is incorrect. Mostly because I am not always watching my graph to see when something is wrong
Ethan, thanks for the feedback! You’re absolutely right. We want to allow users to be able to give detection feedback in other places (like your timeline) and are looking at adding this functionality in the future.