How Accurate is Sense? Summary of Sense Accuracy Measurements vs Utilities

Here’s a quick overview of many Sense measurement comparisons vs. different utilities’ metering systems. In all of these, Sense has proven to be quite accurate, though not necessarily at first. A number of them started out with large errors (the ones with the word “after” in their description), but eventually yielded very positive comparisons. If you are having trouble seeing similar accuracy, some of these examples might help you sort through your differences between Sense and your utility. Issues might include a second panel that Sense isn’t measuring, measurement interval differences between Sense and your utility, or your utility giving inaccurate readings (really !).

Measured Accuracy URL
Within 1% of utility after discovering mystery second panel Usage Accuracy
Within 1% of utility, even after Sense issues, far better more recently How to Tell When Your Sense is Lying to You - #6 by kevin1
Variance of 3% or less vs utility Total kwh accuracy? - #9 by peter
Within 1% of utility after the Sense and utility time periods were aligned Solar Production Comparison
Better than within 1% of utility How Accurate is Sense vs. Utility Metering? - #6 by Dcdyer
Better than within 1% of utility, but only after comparing against utility Actual Reads vs Estimated Correlating NVEnergy Usage Data with Sense Data
Better than within 1% of utility Sense accuracy compared to meter - #4 by samwooly1
Within 3%, in just a spot check against the meter How Accurate is Sense vs. Utility Metering? - #17 by jkish
Within 1%, after adjusting for device breakers upstream from the Sense CTs Electric Bill Question
Within 0.5% - WOW ! How accurate is accurate? (Sense+Solar vs Tampa Electric) **spoiler IT'S VERY ACCURATE**
Very accurate once Sense helped figure out the condo meters were reversed Seeing Big Difference Between Sense and Meter After First Month - What could be causing?
Within 3% on a yearly basis after user figured out the apples-to-apples comparison -had From Grid/To Grid meters How Accurate is Sense? Summary of Sense Accuracy Measurements vs Utilities - #65 by blamb.kt
Very accurate once open CT was closed from Reddit here

If you want to try checking accuracy yourself, here are some tips and techniques.

ps: Let me know if I have missed any interesting measurement comparison stories. Thinking we might need a separate table for solar output comparisons.

8 Likes

This is a great idea! All the new users are asking the same questions. There should be more explanations and resources to guide a first time (or an advanced) user.

I agree with adding a second section for Solar.

Thanks for putting this document together.

2 Likes

Just updated for one more high accuracy measurement after a few home discoveries:

2 Likes

I have Sense with CT and Adapters to where 98% of the time
“OTHER” is not seen.

However, I have a ceiling fan that is on 24/7 not detected and not on an adapter and it never shows up in other.

Likewise, Fluorescent Lights in the Laundry Room that never show up in other when turned on.

A 60 watt outdoor flood that never shows up in other when turned on.

Multiple LED Bathroom Fixtures that never appear in other when turned on.

An Apple iOS charger in a wall socket 24.7 that is never in other.

Obviously others as well.

Which leads me to his accurate is Sense? One would think it knows the wattage from the CT on the mains and can add up the rest to see something is missing. :man_shrugging:t3:

EDIT; I can read all the above, yet when a device is on that I can see it’s on - yet Sense does not show any OTHER wattage, obviously something is a miss.

@Beachcomber since the ‘Other’ bubble is just your total detected wattage subtracted from your overall wattage on the mains, if you’re not seeing an ‘Other’ bubble populate that usually signifies an installation issue (something like open CT’s around the mains.)

Double checked. No open CT.

I do see an other bubble about 2% of the time (from unknown sources), but can manually turn on the lights (etc) listed in previous post and OTHER does NOT appear.

This does NOT mean that total wattage stays the same, rising 18-60 watts depending on the device, but no OTHER bubble appears (nor in devices) so it must attribute the higher wattage to a detected/sensored device I assume.

What do you suggest?

Turned on Sous Vide Immersion Cooker about 8:05pm. Rather obvious seeing it cycling off and on heating the water. Plugged into a kitchen outlet that is not on a Sensor. Not detected by Sense either.

However, there is no OTHER bubble.

So where is Sense Allocating the approximate 500 watts :man_shrugging:t3:



Can’t tell exactly when the Bubble snapshot was taken, so hard to sort out by comparing with Power Meter data. There are a couple of non-bug reasons for no Other.

  • A good sized device where Sense has missed an off. Sense still shows the bubble and that load power continues to be subtracted, so Other becomes negative, hence no Other bubble, even when another load is turning on an off.
  • Always On can also remain too large, at least for short periods, and choke off Other.

As can be seen it’s been on and off for hours now with no OTHER in sight.

Simple question.

Where’s Sense allocating the wattage to for the last 4 hours?

Also, based on explanation in last 10 days, we’ve been told ALWAYS ON is a duplicate of other sources. So the OTHER bubble should be there as well, which it isn’t.

Regardless, as can be seen this isn’t always on but heating the water in spurts.

While I can’t speak to why you don’t have any other at all, the fan will not show in other if it is on 24/7 as it will be in the Always On Bubble.
Other = total wattage - detected devices wattage - always on wattage.

This doesn’t seem to the the case in your scenario as you have said you haven’t’ seen an Other bubble in quite some time, but I just wanted to respond to that specific comment about the fan.

To this quote, i’m not sure what thread you are referencing about Always On being a duplicate. It can be duplicated (in visual bubble form only) when you have DCM or smart plugs. But it is not always duplicate data, and in the trends screens, the data is not being counted twice.

As to the accuracy question. The total wattage in the bottom right corner of the bubble screen is very accurate and your overall consumption trends will be very accurate (assuming no data loss due to power cycle or network outage). But yes, sometimes the bubble math or device trends are not super accurate.

1 Like

As stated in the discussion with the last 2 weeks by Sense tech support, anything in the Always On bubble is duplicated from the other bubbles.

Not in total consumption though.

1 Like

@Beachcomber Could you submit a ticket to Support?
All of the reasons that I can think of for a missing Other bubble have been highlighted in the comments above and I’d like to follow-up with a Support lead once they receive your ticket.

I haven’t seen you conversations with Support, but I have seen them give incorrect information in the past, or information that is correct for the specific case they are working on, but incorrect for the overall usage of the device. I’m just a user, so I can’t say that support is wrong, but 4+ years of Sense plus just looking at my bubble screen right now tells me that what they said is not correct. @JustinAtSense may be able to clear it up, but to add in why I feel support is wrong or the statement was misunderstood.

As I said in my first reply, some data is displayed as duplicate values. Some is not.
If you have Smart Plugs or DCM. Those devices will display the total value of power thought the plug or DCM CT’s.

The plug will also calculate an Always On Load if it exists… On the bubble view, Always On will include the Always on from the Plug.
When you tap the bubble for a DCM device or Smart Plug device, the bubble will show the always on for that.
These two places are the duplicate values. If you try to total all the wattages displayed on the bubble view and compare it to the total live power consumption in the bottom left, it won’t match because the Always On is being displayed twice.

If you tap on the Always On bubble to look at its trend, you will see that it breaks down Always On into “Estimated unknown” and detected Always On. The unknown inside of of Always on has nothign to do with the Unknown bubble. Its just using the same word.
If you look at a power trend for Always On for the month, it includes the always on power from Smart Plug Devices.
If you look at a Power Trend for a smart plug device for a month, it includes its Always on. So again, if you try to add these things by hand, you will get incorrect numbers because the data is displayed twice. But when coming up with total consumption numbers for the month, Since Sense knows what is what, it is not doubling anything as the monthly total consumption for the unit is not calculated based on the sum of all the devices, but the actual power that crosses the main CT.

The non duplicate data -
If you have Always on power that is not via a smart plug, this is only shown in the Always on Bubble.
Unknown = anything that is not detected. Always On counts as a “detected” device.
See the two screenshots from my Sense as of a few minutes ago.
Unknown is 184w.
Always on is 362 watts
This is correct as I can tell you exactly wha is and is not on right now.
The 184w is a fan that I turned on to make an Unknown bubble appear on the screen.
Otherwise there was no unknown because its quiet in the house right now and everything else is detected between being captured by Always On, smart plugs or native detections.


I think you mean the reverse. There can be an Always On component from each DCM and smart-plug, that is duplicated in the house level calculation for Always On. That’s one of the two reasons Always On can be larger than it should be. The other is the time history of the house level Always On. It reflects a near minimum of what happened in the house over the past 48 hours, but if the house number is headed lower, it can be too big to reflect the current state of usage, and takes some time to catch up. It would be interesting to see how big the sum of DCM/smart plug Always Ons are. You can see that in the Alway On device details.

Sense doesn’t really allocate a new undetected device to Other. It just does math to calculate Other: Other = Total Usage - All ON Devices Usage (including Always On). When that math goes negative, you won’t see a Bubble for Other.

I was hoping to see a couple of bubble shots with time stamps that aligned with places on your power meter, just to see is any bubble appeared stuck on. But if you haven’t seen an Other bubble for 4 days, it’s not likely to be a stuck on bubble, so, I would contact support as Justin suggested.

The way I think about it is, not seeing an “Other” bubble is usually a good thing.
Not seeing the Other bubble mean your undetected devices are currently using the same or less power than your average over the past few days.
“Always on” is is an average wattage for all the unknowns. “Other” reflects anything that is above that average.
Other is another way of saying the amount of unknown power you are currently using above your average usage.

Average meaning, average of everything that has NOT been detected over the course of days. Average is not real-time per say.

When you add in the consumption of all your detected devices you could very well be using more overall power.
That is where the “trends” data comes in.
Using the mobile app you can see a breakdown (percentage) of all of your devices. This is where you can really see your biggest power hogs and get the information you need to get a grasp on where and how you can start saving energy and lower your bills. Why Sense doesn’t provide this percentage breakdown in the Web App along with some of the other most useful tools Sense offers is (no offence to those with no control) assinine.

If you never see an other bubble including for a brief second when a device first turns on before it’s recognized as a discovered device, something is wrong. Try turning on an undiscovered device that uses a bit of power and see what happens. But don’t expect something that uses little power or is on most or all of the time to appear as other.

Reading the whole thread again @Beachcomber is it possible Sense is confusing your AC and your cooker. Another possibility is that Sense may be missing your AC’s off signature thus attributing any wattage that would otherwise be unknown to your AC.
I see that your air handler is also on, so I can only assume AC is the AC compressor itself. It appears that it uses about 550-560w which is about the same as your cooker. While unusual for Sense to confuse a motor and a heater it’s quite possible. Sense might be confusing itself thinking your AC is on twice and doesn’t know how to report it. Guessing by the size of your AC bubble reporting double your actual AC usage by including the cooker wattage.
When your AC is off turn your cooker on and see if you AC bubble shows up.

1 Like

My Air Handler is not detected by Sense natively. My 2 additional CTs are monitoring that circuit which is the only way it could be identified.

So there is 0% chance that Sense attributed the 120v 500 watts from the Inversion cooker to that dedicated monitored 240v circuit.

1 Like

@Beachcomber, are you still bereft of an Other bubble all the time ?

99.9% of the time I am missing an other bubble even when I have known lights or other wall warts that are on.

It even misses my 240V Dryer most of the time.

But of course as I type this, it’s actually showing.


I do notice looking at the bubbles there is a lag time - in other words it’s not in real time producing the (correct) bubble. For example, sometimes when the AC Compressor kicks in (or turns off) the Bubble lags significantly.