The third-generation Oura Ring will receive an updated sleep algorithm; the activation of the spO2 sensor could conceivably change some sleep data calculation – but both are not yet here.

So, sleep data should currently be similar or even the same..

In this post, let’s check out the sleep score, sleep stages (phases) analysis and related data.

Sleep Score

Sleep (and with that, recovery) has been the original focus of the Oura Ring; it should analyze that well.

Let’s start with the sleep score that summarizes everything.

Sleep Score, Oura Ring Gen2 vs. Gen3
Sleep Score

In the sleep score, as with readiness, there are some differences.

Except for the two values towards the end of this comparison period, when the Gen3 Oura ring shows my sleep score as trending upwards while the Gen2 found it to be declining, the values and trends are similar (if not the same).

Sleep, I have come to find, is interesting because a closer look shows what data is based on very strong signals, and what is harder to determine.

The best example of strong signals and easier data: bed and wake-up times.

Bedtime

For a wearable to recognize when its wearer has gone to bed must be relatively difficult.

When someone goes to bed to read and stays still or chills on a couch watching a movie before quickly falling into bed, the shift from that to sleep is probably not so easy to catch (and that’s also related to latency, the time between going to bed and falling asleep, which is yet another measure).

I think I made things rather easy for the ring(s).

Sometimes, I may have read for a bit or given my wife a massage, but the switch from activity to bedtime is quite clear – and that may well show here.

Bedtime, Oura Ring Gen2 vs. Gen3
Bedtime

The Gen2 ring regularly saw me in bed a bit earlier than the Gen3 (maybe when I was still moving my right hand turning pages or giving said massage), but the agreement between devices is very good.

Latency

Since it’s related, let’s check the latency next:

Sleep Latency, Oura Ring Gen2 vs. Gen3
Sleep Latency

And we see, as I said, that the switch from lying in bed to actually falling asleep might be harder to measure than that from activity to bedtime.

This is some of the sleep-related data from the rings with the most disagreement.

There are differences of up to 13 minutes in how long the Gen2 versus the Gen3 ring “thought” it took me to fall asleep.

This may be another effect of my having worn the Gen2 ring on my non-dominant hand, and of it having thought that I was in bed earlier than the Gen3 ring saw the relevant signal of that.

The most divergent latency data was not, however, when the rings disagreed about my bedtime, so the answer is not as simple as that!

Wake-Up Time

Waking up is an even more obvious signal than my falling into bed.

It’s time, I’m up and start to get active. No snoozing, no prevarication, no tossing and turning.

That “up-and-go” should show, and it does show rather well.

Wake-up Time, Oura Ring Gen2 vs. Gen3
Wake-up Time

There are some times with different wake-up times measured by the two rings.

That’s probably when I woke up a bit too early (let alone, had to take the dog that sleeps in our bed out), then slept for a bit longer yet – and the Gen3 ring decided to use the earlier time as my wake-up time.

These (more-widely) different times are rare, however.

Most of the time, the wake-up times measured were pretty much the same.

Total Sleep

Between bedtime (well, plus latency) and wake-up time, minus awake times, lies the amount of sleep I got in total…

Total Sleep, Oura Ring Gen2 vs. Gen3
Total Sleep

… and the graphical representation might not be ideal; the scale makes the differences look small.

The differences are never too bad, however; around 15 minutes in a total time of around 7 hours isn’t horrible.

Sleep scores could be different because of a difference like that, but as we’ve already seen, the sleep scores are comparable.

Sleep Efficiency

Sleep efficiency is a measure of how much of the time in bed a person was actually asleep.

The measure is… let’s say odd, looking at the comparison.

Sleep Efficiency, Oura Ring Gen2 vs. Gen3
Sleep Efficiency

It’s not completely random, but there is a lot of disagreement.

It’s not the greatest differences all the time, but a sleep efficiency that could have been either 73% or 89%… well, that *is* a difference, especially as 85% is the cut-off for good sleep efficiency.

Total sleep sounds like it should be the major contributor to sleep efficiency, but the relationship between those two measures is not very close.

Restfulness

Another measure related to sleep quality is the restfulness, the tossing and turning during sleep.

Restfulness, Oura Ring Gen2 vs. Gen3
Restfulness

There is quite some disagreement here, too, but not quite as much as with sleep efficiency.

Values and trends are rather more similar. Outliers to exist, however.

It all makes me wonder how much I move my left vs. right hand when I sleep; that would be one potential explanation–if a pretty strange one.

Time in Sleep Phases / Sleep Stages

How long I was in different phases of sleep, that’s something where the Oura ring comparison does show lots of differences.

It’s the same, of course, if we display the sleep separated into the relative amount of time spent in its different stages:

Sleep Phases, Oura Ring Gen3
Sleep Phases, Oura Ring Gen3
Sleep Phases, Oura Ring Gen2
Sleep Phases, Oura Ring Gen2

Honestly, I am not sure what to make of the differences here.

They are noticeable sometimes, but one has to be careful to check the scales again, as they exaggerate some of the durations somewhat.

However, they are never too bad, once again.

Mainly, I am thinking that a close look at these results shows that a close look at these results might not be the best idea.

The differences lack context in and of themselves (why did they come about?), and there is little reference by which it would be possible to argue that one or the other of the rings was more accurate.

If I had a sleep lab or one of their (EEG, at least) monitors, that would say something.

This being as it is, it’s probably more sensible to just look at the sleep score and the advice given by the Oura app rather than cause sleepless nights wondering what exactly the sleep phases might tell.

Also, Oura Ring Sleep Algorithm Update…

There will be a new, improved sleep (stages) algorithm in the third generation of the Oura ring. As so many of the new features, it is only scheduled to come later in 2022, but still: The above data should just be what gets analyzed so far; this will change again.

If you still don’t have enough of it, however, I will soon publish a week of (this) comparison data with a Fitbit Charge 5 added to the mix!