Fitbit sleep phases inaccurate

[veloron2008]

New Fitbit Alta HR user here. 

I've had two very restful nights of sleep in a row, and the Fitbit tracker recorded <1 hour in deep phase each time out of 7:30 total sleep. I know this is false for a few reasons:

First, I subjectively feel very fresh and rested upon waking. No drowsiness whatsoever. Second, I also am still using my old trackers for comparison (a Misfit Ray and a Mi Band 2), both which indicated more than 35% of total in deep sleep. I also recall several dreams which makes me skeptical of such a high % of light sleep. I am a vivid dreamer, especially since supplementing with Magnesium at night. Finally, I awoke at least once last night because of tingly hands due to staying in one position too long (side sleeper). That's an indication of deep sleep. 

So now I'm reading that HR-based Fitbits use heart rate fluctuation in their calculations. Since I'm an endurance athlete with a resting HR in the low 40s, a few beats here and there is pretty significant. 

I would say to other Fitbit users to take the sleep analysis with a HUGE grain of salt. 

[WavyDavey]

---

@veloron2008 wrote:

New Fitbit Alta HR user here. 

I've had two very restful nights of sleep in a row, and the Fitbit tracker recorded <1 hour in deep phase each time out of 7:30 total sleep. I know this is false for a few reasons:

First, I subjectively feel very fresh and rested upon waking. No drowsiness whatsoever. Second, I also am still using my old trackers for comparison (a Misfit Ray and a Mi Band 2), both which indicated more than 35% of total in deep sleep. I also recall several dreams which makes me skeptical of such a high % of light sleep. I am a vivid dreamer, especially since supplementing with Magnesium at night. Finally, I awoke at least once last night because of tingly hands due to staying in one position too long (side sleeper). That's an indication of deep sleep. 

So now I'm reading that HR-based Fitbits use heart rate fluctuation in their calculations. Since I'm an endurance athlete with a resting HR in the low 40s, a few beats here and there is pretty significant. 

I would say to other Fitbit users to take the sleep analysis with a HUGE grain of salt. 

---

Hi @veloron2008 ,

A few thoughts:

The Fitbit number sounds more correct, sorry. I somewhat doubt you're getting 35% deep sleep nightly--that's double the average and would really cut into other stages. That you feel rested is great, but doesn't provide any information as to the cause of it being a huge amount of deep sleep. (In the time I've been here, I've never heard of deep sleep being reported so high). Here is a more typical chart:

Often, we see the light sleep % higher, usually at the cost of deep sleep.

  It's very easy to take our subjective experience, and try to link it to the sleep stage numbers and call that objective. But there are a lot of assumptions to doing that. A scientific study would be more accurate, at the cost of quite a bit of money. I think Fitbit's solution is a good value overall for what it is.

[veloron2008]

To clarify, my other trackers do not attempt to distinguish Deep Sleep from REM Sleep, so the ~35% includes both phases. According to your figure, that benchmark is actually 40% not 18%. 

Fitbit puts my Deep Sleep plus REM at 18% of the total, which is absurd given the facts in my first post. 

I maintain that the Fitbit algorithm isn't accurate, sorry. 

Can you give some details how Fitbit uses HR to calculate sleep phases? Are fluctuations not more impactful to an algorithm for lower heart rates, than for higher heart rates?