How does Fitbit calculate Resting Heart Rate?

[BigHaole]

I'm finding this very frustrating to be sitting at my desk, and see my HR showing as 74, but my Resting HR being 80.  Clearly Fitbit is not using the conventional definition (from Wikipedia):

"The basal or resting heart rate (HRrest) is defined as the heart rate when a person is awake, in a neutrally temperate environment, and has not undergone any recent exertion or stimulation, such as stress or surprise."

This definition would lead me to expect my reported resting heart rate to be the low value reached in the early morning, or at least the low value I reach, during the day.  Instead, it's above both of these.  I would like to know how it is being calculated, so I can know if my Fitbit is reporting anything useful when this number goes up or down.  Over the recent new years holiday, I got more sleep and more exercise, with less stress, so I was expecting this to go down, but it has gone up and I do not understand why.

[FidRewe]

Can't believe, fitbit hasn't managed to draw appropriate consequences from this discussion for five years. Why is fitbit doing that? It would be so easy to get an accurate resting heart from the time spend asleep! Of course, it's the same in my data... the heart rate during inactive periods during the day will be higher than the resting heart rate during the night, so the measured "resting heart rate" will always lie above. It would be fine, if both heartrates were determined in parallel, but if an average is formed over resting periods in general, the measured "resting heart rate" will depend on the duration of sleep and the amount of time spend not moving during the day. Which destroys the validity of the parameter.

Is anyone aware of an app that overcomes this unnecessary insufficiency?

[AquamanJon]

So here is a data point. This morning, my resting heart rate was 60ish. The app showed it as 70. I got out for a run in the heat and I average about 150BPM for 2 miles, I sit down. Look at the testing HR. It’s up a point to 71.

This supports the overall daily average theory recently advanced.

Moderator edit: personal info removed

[jonengstrom]

They RHR is a mess - it doesn't need to be. They've got millions of users, why not make the best of it.

I'd be happy just to see my sleeping heart rate, in order to e.g. see if I had an infection or whether my heart rate changed over time. Looking at my RHR history gives me 0 value.

[notthatfast]

I've found on several occasions my reported RHR is simply the average of my lowest nocturnal HR and the lowest non active day time HR.  Use your Heart Rate display screen to test this theory and see if it works for you.

[Oregonmama]

Not even close.

[Danoxonuk]

I agree BretAllen but I’d say it’s a rolling average across several previous days. Drink alcohol four evenings in a row and see the rhr gradually climb. It doesn’t spike day 2. Don’t drink the next four and watch it gradually descend. It doesn’t drop back to normal straight away. It’s smoothed. It’s a rolling average over a number of days.

[Oregonmama]

If it were the average of the lowest day and night rates, it could not be
higher than the maximum daytime low for the past week. Mine is, on a very
consistent basis.

[UtrechtNL]

Okay, I haven't contributed to this discussion yet but I have been closely looking into the heartrate values, looking at my approx average value during the day, comparing to my sleeping HR (when I got the premium trial), etc.

My conclusion is that a lot of it seems to be calculated during your sleep, but there is a huge moving average that is probably implemented the wrong way, and biased towards detecting upward trends vs downward trends.

So for example my RHR would be around 65. Then one evening I drank a pretty large amount of alcohol and had a relatively bad night:

Day 1: 65 - nothing up yet

Day 2: 65 - drank quite some alcohol in the evening, HR went up to 85 or so for the whole evening, no effect on RHR despite

Day 3: 67 - night with relatively high HR, immediately a +2 bump

Day 4: 69 - despite the night + day 3 was pretty normal again, it would go up even more

Seems like indeed it's a moving average, and either spikes get carried into the future, or there is a smoothing algo that prevents it from changing too much in 1 day. E.g. max 2 per day (though I've seen it hit +3 at times).

Also after exercising very hard for one day, it still raises the RHR while it's not supposed to do that because it shouldn't use values after exercise for that (+ possibly lingering slightly higher RH because of recovery after).

So, my life has been pretty stressful for a while, but the problem with the Fitbit algo is that it enlarges any abnormality to the point where it's no longer possible to use the value in a useful way because one super stressy day affects 3 days of RHR...

All of this isn't true for the sleeping HR, that you can't see without paying for premium; that one is actually just a per-night value. To me that one gave a pretty good indication of actual stress levels. The RHR is indicative of that too, but its resolution is just god awfully bad because of all the moving average or enlarging of any detected trends in its change. Plus exercise and alcohol have a big impact on it so that makes it less useful.

So in conclusion: the Fitbit RHR is not really a useful value for people that sometimes drink alcohol and have irregular exercise regimes. Or have nightmares that wake you up with a high HR. The RHR will bounce all around because of the heavy algorithmic stuff around it. Sleeping HR is much more useful but unfortunately you have to pay $11 per month which is just not worth the tiny amount of extra data you get @ Fitbit. 

I feel like Fitbit needs you to lead an exemplary life for it to give you a correct RHR, or you just have to pay to see your sleeping HR which it otherwise keeps from you. For the $11 you get your sleeping HR and a useless "tossing and turning %" which to me didn't seem to correlate with how my mind or my body felt at all.

Btw Garmin is pretty transparant on using sleeping HR to generate RHR, and using a rolling window of 7 days for some devices, though they haven't disclosed their full algo either. Looking to maybe changing to a more serious device from Garmin to get some useful stats that I don't have to pay $130+ per year for.

[OldRattler]

I'm still getting used to the Fitbit Versa 2 compared to the Garmin Vivosmart 3 it replaces. The Versa 2 is a better watch but I've been surprised and disappointed by the app, and its dependence on external servers. The Vivosmart was quite capable of reporting 7-day average resting heartrate on its own, with no connection to the app or the Garmin network. However, the calculation was equally mysterious. Overall the Garmin was more forthcoming with basic data than the Fitbit Versa 2. The Garmin was also more accurate at measuring steps.

[TEvans1]

I understand proprietary intellectual property, so I don't expect full disclosure but I'm curious.  My average sleeping heart rate last night was 57 and my range was 48 to 72 with approximately 4/5 below 61, (my resting HR in the app).  If 82% is below resting heart rate and I just came off two straight days of 59 RHR, I don't understand how/why it went up.

Thanks for any help you can give.  I've owned a fitbit for a long time and this inconsistent and somewhat ambiguous resting HR is making me consider going over to a Garmin.

[gateswood]

I've been using a fitbit since 2015. My RHR graph on the fitbit.com dashboard always looks like the Rocky Mountains or a snaggletoothed saw blade. Always.

If you want to know what's really going on, use one of the API's to view the raw data. The fitbit logs a HR reading every 5 seconds or so.

[Metsci]

Can you teach us how to access the API?

[gateswood]

Sure. My pleasure.

One is at https://www.squashleagues.org/ Register, log in, select Fitbit users up top and go explore.

Another is at https://iccir919.github.io/pulseWatch/public/index.html. No account needed. Just sign into fitbit and allow it to access your choice of data.

At either one you can download your data in csv files or view it graphically. The second one has a much better graphing functionality.

[Mary48]

It's goofy all right. One time when i was walking, the HR on my fitbit showed 160, when I physically check my pulse, it's normal. I did not feel the palpitation, believe me, at my age, if its 160,I'd know it. Then one time it said i was in aerobic mode in the middle of the night but my HR is 65. My cardiologist thinks fitness trackers ate not 100% accurate.

[dlhnj]

I would agree that the tracker is not always accurate during exercise.  At rest I have tested the tracker often and it is very accurate.  But when walking I sometimes get spikes that are very high, lasting for a few minutes, that are clearly not accurate. 

[BretAllen]

I have been using fitbit since 2015 and I am convinced that the resting heart rate is simply your AVERAGE heart rate over the day.

[pesegers]

So, not "resting hart rate" but "sleeping hart rate". Fitbit is cheating on us!

[pesegers]

Yes, i have comments. How exactly do you calculate the resting hart rate? What is the formula? Fitbit pretents to be of help with covid-19. But you can't even clear out a simple question. How exactly do you calculate resting hart rate? No sweet talking please. The formula.

[Lostncspace]

The secret algorithm does not agree with the authoritative definition of resting heart rate. This is a non-answer.

[AndrewMir]

• You guys need an upload screenshot/attachment button for your mobile site
• second point is that I woke up and checked my sleep heart rate, it uploads, calculates it high, then finally minutes later adjusts it
• lastly, my rhr seems high according to the Fitbit readings, like I’m sitting down with a hr of 55-58 (Never mind low of 52 while sleeping) and it’s saying my rhr is 60 lol.