Sleep score for short sleepers

I only need 6 hours of sleep per night, but my new Garmin Epix 2 keeps telling me that my sleep quality is "Fair" due to the number of hours of sleep, even though my other sleep metrics (like deep sleep, REM sleep, etc) are all excellent. Two questions:

  • Is there any way to manually adjust the number of hours of sleep per night that the algorithm expects? I noticed that I can override several of the defaults, like heart rate zones for example. Is there a way to do this with sleep?
  • How does the sleep quality affect other calculations, especially the recovery time metric? I just did a 45 minute run and it is telling me 72 hours recovery time. I feel great and ready for another run. Is the recovery time affected by my low sleep duration, or is it due to other factors like my HR during my run?

Some background. Most people require 7-9 hours of sleep per night to function optimally, but not everyone. Sleep needs per person fit into a normal distribution (bell curve). The 7-9 hour recommendation is based on what 95% of us require (2 standard deviations), which is typical for most statistics-based recommendations. But 5% of us fall outside the normal range: 2.5% require less than 7 hours and 2.5% require more than 9.

I only need about 6 hours of sleep. I had this confirmed by an MD sleep specialist who who treated me for insomnia. One of the recommendations was to stop trying to sleep so long. If I succeed in sleeping too much one night, then my body reaches stasis by not allowing me to sleep the next night. For example, on the occasions that I am able to sleep 8 hours, the next night I will wake up after 4 hours. It doesn't matter what time I go to sleep or if I exercise, relax before bed, stay off devices, avoid alcohol and caffeine, etc. Those are all great and it's important for me to follow them just as much as everyone else does, but they don't make me sleep longer.

Sleep researchers have observed that people who need less sleep tend to get the same amount of deep sleep and REM sleep as everyone else. They just get less light sleep in between. This appears to be the case for me, according to my Epix 2. I attached some screenshots from last night. All nights look about the same as this one, since I started wearing the watch a week or two ago. All of my "Sleep Score Factors" are "Excellent", except for "Duration" which is "Fair"

Is there any way to customize the Epix 2 for users with unusual biological sleep requirements?

I really enjoy the sleep tracking, and it has really encouraged me to practice good sleep hygiene. I haven't tried bad sleep hygiene since getting the watch, but I am sure that whenever I used to use devices before bed and all that stuff, my deep and REM sleep suffered and I woke up feeling exhausted. I am already sleeping better because of the Garmin, even though the sleep score metric is inaccurate. It's just that I worry that other metrics like recovery time might also be inaccurate.

Thanks,
Scott

  • Regarding your questions:

    1. Is there any way to manually adjust the number of hours of sleep per night that the algorithm expects?

    No, and you really do not need that. Just read along.

    2. How does the sleep quality affect other calculations, especially the recovery time metric?

    It doesn't affect other calculations much.

    Now, I am in a slightly worse situation regarding sleep score, my average over the last couple of months has been around 65. In time I realized I am an exception, and do not expect the watch to have a well-defined rule specifically for such cases.

    Nevertheless, the recovery time is mainly (or even exclusively, based on my observations) based on the effort intensity, taking into consideration your age, fitness level, and heart rate zones. E.g. after a 1.5 hour run workout in HR2 the watch recommends me about 20+ hours of rest, but if I do a 30 minute workout in HR4-5 it recommends much more recovery time. That is contrary to my perceived effort, I am following a long term low carb diet as I have type 2 diabetes, and because of that I feel more exhausted after a long and easy workout than after a short but hard one, no matter the difference in intensity. But the higher suggested recovery time takes into consideration the strain on your muscles and joints as well, which you may not feel immediately, but it can lead to injuries if you keep it hard for a longer time. Most of the times you do not feel when you're overtraining.

    My advice is to follow the suggested run workouts (even if you may find them too easy) and recovery times on the watch for few months and notice the results afterwards. They are very well conceived and helped me improve my VO2 max by 2 points in 9 months, which I am very happy about. I also did benchmark runs every 3 months and was very pleased to notice the improvements; meanwhile I stayed injury-free which I hardly believed possible - as I have had a problematic knee joint for years as result of an old injury.

    And I would say not to worry much about sleep scores on the watch, as long as you are healthy. It just knows what it is doing regarding your fitness.

    Just observe and learn the watch so well that you can adjust your expectations based on its potential, make use of its strengths and live with its weaknesses.

    Hope it helps, good luck.

  • will concur with this considering I am a 4-6hr sleeping (recovering from chronic fatigue due to sleep issues so when it gets deep sleep I am like an energizer bunny). While its going to impact the metrics slightly its not enough to overly impact them - training readiness for example will probably never be a glowing purple for example, but you work it off yoru baseline i.e. green instead.

    Recovery time is more about when your body is ready to give max effort - doesn't mean you can't and shouldn't exercise - but may mean its a better option to do a nice recovery benefit activity. Recovery time will adjust based especially on Stress (so if body is tired and maxed out this will reflect in your stress - if it really wasn't a hard workout and watch is still adjusting to your baseline then it will pick up that your stress is low and recovery time will reduce quicker. You can improve this by performing a recovery benefit exercise as when it sees load (EPOC) etc from this it will adjust again taking that and the prior exercise into account. Sleep factors a bit - but its more from the stress part of sleep longer sleep and stress is excellent - more/quicker recovery - if you sleep is short there is less excellent stress (i.e. stress free) so less chance to irmprove recovery time.

  • I ended up deciding to take the watch off at night and disable all sleep-related screens from the display.

    Since the sleep score didn't affect the other calculations, there wasn't really any point to using it. It was just a number that I could fixate on. I kept trying to make sleep score 100, which wasn't going to be possible, so it was stressing me out for no benefit.

    The New York times just wrote an article about this. It's not just Garmin. All sleep tracking devices are pretty much the same in this regard. They fail to account for a variety of individual biological differences, they're based on incomplete science, and the sleep scores seem to motivate some people while causing anxiety for others.

    Three months later, I'm still trying to figure out precisely how a workout might affect my recovery time. It's pretty important, since I often have plans later in the week and I want to be in top shape for some activities whereas others are just training time. How hard I plan to work out in the near future isn't integrated into the recovery time model. It's definitely true that harder workouts result in much longer recovery times. In addition to that, it seems to give a "credit" when the baseline activity level over the last few weeks was already high to begin with, and it seems to give a recovery-time "penalty" for a rapid increase in activity level. But just like with the sleep score, I can always ignore the recovery time number and go with how I feel.

  • Having a low(er) sleep score or not having a sleep score will only affect your training readiness score. With missing training readiness metrics, you cannot max out your training readiness ("prime level"), but you can get it in the green zone ("moderate" ie "good to go")

    www.garmin.com/.../

    Your recovery time is influenced by the (lack of) quality of your rest/sleep, level of stress but not by the sleep score per se. Poor sleep will delay recovery a bit, and great sleep will accelerate it a bit. The recovery time is based on EPOC, mostly driven the intensity and length of your activities.

    Just like any wrist based device, the Epix is not great at tracking sleep phases or even awake/sleep state. In ideal situation, when using an ECG device to capture data (not the wrist based HR sensor), the Firstbeat algorithm will identify the correct sleep state less than 2 times out of 3, with the lowest score for REM and awake time. So don't put too much faith in your sleep score anyway. You can also see in the matrix below that the watch "light" state will in fact be confused with another state with equal chance. Now, if you take into account the challenges of wrist based sensors, these numbers will be lower in practice.

    How to use the sleep tracking then? A bit like the HRV status. If you train too much, your sleep quality will suffer, and mostly because of the imbalances in the autonomic system due to stress. I personally have been using the sleep score with good success to track the benefits of CPAP and fair success to optimize the settings of my CPAP.

  • It should never be about getting 100% - its should always be about the trend and your improvement over time.

    If you were to run a 100m, would you expect to get under 10s? That is say garmin's 100%. No you woulid track you time and say it was 13s. So that would be your base - your focus would be to then improve on tthat 13s.

    That is effectively what all these metrics are about. Sure if you are getting 100% straight off the bat for sleep tracking, great, but more important would be to be consistently within your range, possibly imporving. For example my range is 60-75% - if in there I am doing great (especially considering I can only get 5-6hrs sleep max), if under 60% I am looking at the factors as to why so bad, same as if over 75% what did I do differently - maybe I managed to get 6.5hrs for a change.