Sleep tracking accuracy ios16

It seems with this change Auto Sleep just reads and displays the IOS data, ignoring it's own.

CapeTownMacaroni wrote:

I did this and now Autosleep and the Apple algorithm agree but the deep sleep value is still way below what it was in the past.

Hey CapeTownMacaroni,

I am no sleep expert, but have experienced similar "scoring" issues comparing Withings Sleep Mat, Apple Watch and other random wearables since I became interested in health tracking.

I have no association, but if you are interested in digging deep on this I would take a look at The Quantified Scientist. Tons of science based information, comparisons and reviews.

Hope this helps.

All the best.

Interesting conversation Cape Town,

I cannot argue the merits of the Apple Watch Algorithm. If for no other reason than I do not have the knowledge or a large enough personal sample size to produce any data with significance.

TQS does a great job with what he does, but as you mention, his test is on himself which is a baseline of one and does not consider the cofactors associated with other individuals or groups.

Personally, I consider the Apple Watch in the top ten of consumer devices that can be purchased easily, over the counter, at an affordable price to monitor sleep or other health metrics. That being said, I do not consider it a medical device (as close as it may get at times).

Your sample size of 15 with your questionnaire affirms some of my assumptions that that Aura ring is probably better suited at sleep tracking. I would also make the assumption that the sample is a little bias as those usually associated with a biohacking group are more healthy, fit or aware of a health lifestyle than the general population, which may sway the results versus a larger, general population sample.

As a last ditch effort to assist I would do the following if you have not already:

https://www.myhealthyapple.com/apple-watch-not-tracking-sleep-or-sleep-stages-heres-what-you-can-do/

I am not sure if you have stumbled upon this conversation:

https://www.reddit.com/r/AppleWatch/comments/xrxkfw/am_i_the_only_one_whos_aw_records_very_little/

I enjoy the conversation, but I really have nothing else substantial to provide towards a solution as related to the Apple Watch (I am not sure an Genius is going to get you anywhere with this either). I would find it interesting if you were to track with a different device (Aura, FitBit, etc.) if your results would match your current conclusions or create another outlier?

If you find the time, post back any results or solutions you may find (or if you happen to find an ear at Apple to assist with this).

Good luck. All the best.

I wear a Fitbit Inspire 2 on the same arm next to my Apple watch 6 and I also get lower deep sleep on the watch vs Fitbit. The REM cycles dont match either. The values can be off 2x-3x. I dont buy into the studies showing apple watch is the most accurate wearable since my blood oxygen readings can go from 100% to 85% in the same day. My skin is fair with no tattoos. I tend to believe the fitbit more since deep sleep is a key factor to me and i know what 1+ hrs does to me vs how i would feel with the fewer minutes Apple watch shows.

Never get 5 hours of deep sleep with auto sleep.

Apple Watch consistently underreports deep sleep and it is very strange data.

I have a recent garmin fenix 7s and the sleep data is much more reliable.

Come on, guys! What do you expect? Sleep studies attach wires to your head, your arms, your feet, and your chest! They measure actual eye movement, multiple aspects of heart signals, and alpha brain waves! The amazing thing is that an Apple Watch or a Fitbit or whatever can work at all!

You can't compare the Apple Watch data to the Fitbit data, and certainly not to real sleep study data. What you can do is monitor your sleep each day and be aware of drastic changes that might indicate something's going on. This ain't brain surgery, and you don't want your brain surgeon depending on either an Apple Watch or a Fitbit. Let's use it for what it's good for and not complain about what we have no right to expect.

I think you are missing the point. I am not expecting exact wires to head data. Those of us that have used FB for a long time have gained a correlation confidence from what we feel from sleep and the data. The data from the Apple watch many times just doesnt match how we have experienced that sleep for whatever reason. My expectation is that the two should be somewhat in the ballpark but not dramatically different.

Maybe if we can get the algorithms of each that determine what the states are, we can understand where/why they diverge at times.

Hi all,

is this not all down to the simple fact that AutoSleep’s Deep Sleep metric is in fact just the Deep and REM stages from Apple Sleep added together?

And doesn’t that mean that AutoSleep is misleading in its reporting of ‘Deep Sleep’ in the clinical sense of the term?

EDIT: On second look, this appears to be the case for some of my results but not all of them. What I don’t understand, though, is what AS Deep Sleep actually represents. Is it the actual deep sleep stage, or some proprietary metric that they’ve come up with that doesn’t really compare to anything else?

I have been using Fitbit for years and just switched to Apple Watch. I was consistently getting 13-18% deep sleep on fitbit but get only about 1%. That cannot be correct. I feel refreshed when I wake up.

Clearly there is a disconnect between the criteria defining "deep sleep" of the various entities (AutoSleep, FitBIt, Apple, et al).

The best way to eliminate the confusion for end users is for the different groups to publish the criteria they are using (based on agreed parameters). This would also include the measurements made by the various devices, the expected accuracy of the device measurements, etc.

Face it -- rubber rulers don't work. Standards are needed. . .

I’m observing the same thing with my Apple Watch 7. It concerned me enough that I did a search to see if others like yourself have had the same experience. I have been logging 30 min on average for deep sleep. This is well below the recommended.

Before my Apple Watch, I had a Fitbit Ionic for years. It consistently recorded that I was getting the normal level deep sleep minutes.

I’d like to think the Apple Watch algorithm is off. That said, I don’t often wake up well rested even on a solid 8 hrs. And so it has me thinking that maybe the Apple Watch is picking something up on my sleep patterns that the Ionic did not?

I can support your concern with Deep Sleep recording . In 2020 using my wife's Fitbit I was clocking between 5 - 16% Deep Sleep . I bought Apple 8 series watch about 8 days ago and its recording Zero Deep Sleep most nights !! . My wife and swapped and she recorded 7 mins deep using my watch , but there were some inconsistencies from her watch that I wore . Ive had disturbed sleep patterns for some years , which match REM Sleep Behaviour Disorder symptoms which was my reason for buying the watch. Any research on quality of sleep details the importance of Deep Sleep both in quality and quantity and the rather scary implications should you fall short on this. We are going to try more swap tests and both of us using both watches to compare . It was something of a comfort to hear Im not alone with concerns about Deep Sleep accuracy on Apple Watches. Richnsoul

UK

Thanks I have done as you suggested as Apple Health deep sleep figures correlate much better with how I feel I have slept and with the results from Sleep Cycle and more independently from my Wellue oxygen saturation ring results. Sadly as Auto sleep results were far and away better.🙁

Same here. Autosleep for years (though the alarm has an issue with conflicting with Watch alarm so may have to stop) and deep sleep reports at about 2 hours. Apple reports 53 minutes.

FWIW just adding my voice

I’m a long time Autosleep user

Avg Deep sleep 2hr 12

Apple Watch

Avg Deep sleep 18 mins

What’s going on here ?