Whoop vs. Oura Ring vs. Apple Watch for Optimizing Sleep and Workout Performance with Continuous Glucose Monitoring and One Meal a Day: Part 2
In the initial post for this series I shared the basic experimental setup and example data output. Here, in the second and final part of this series, I’m going to dive into notable variables and examine the overall data to conclude initial lessons learned from this 30-day experiment.
As a quick reminder, during the initial part of the spring 2020 COVID-19 quarantine I decided to wear a Continuous Glucose Monitor while testing the differences between Whoop, Apple Watch, and Oura Ring. I had already been doing daily time-window eating between ~5:30-7:30pm (i.e. One Meal a Day, “OMAD”) for 6-7 months, and my ultimate goal with this N=1 experiment was to figure out how to optimize sleep and workout performance by tweaking food/drink intake and timing.
Insight #1: When not in ketosis and doing OMAD, not enough food or too few carbs produces sub-optimal sleep & poor next-day workouts
While in past experiments I’ve noticed that deep ketosis produces high quality (and a tad shorter) sleep and reasonable workouts, not surprisingly in this experiment when I decided to go low carb-ish or skimp on the quantity of food in a meal, the sleep quality (but not necessarily overall length) goes down.
For example, one day I decided to eat early and relatively low-carb-ish (then decided to eat 20 peanut M&Ms to see what would happen, followed by a short & intense workout later):
Interestingly, the 20 peanut M&Ms didn’t really spike blood sugar as much as I thought they would. Also note that light/moderate workouts drop blood sugar, whereas more intense workouts raise it.
Also, the sleep quality wasn’t great, even though the sleep meters said I was asleep for more than 8hrs total.
And in my workout the next day (i.e. a light run followed by light bike), I was totally zapped for energy and couldn’t keep my heart rate up:
Insight #2: A bad night’s sleep leads to more intense blood sugar swings the next day
For example, here is the next day’s glucose readings:
Top glucose levels here read 160 mg/dL, which after calibration was more like 140 mg/dL, FYI (you have to be careful with CGMs).
As an aside, I’ve noticed as a non-diabetic I can’t really get my blood sugar higher than 140-150 mg/dL or so, though doing OMAD certainly keeps my blood sugar higher at night and early morning than when I’m not doing OMAD.
Long-story short regarding meal quantity and composition during OMAD for me; it’s not a good idea to skimp on the carb ratio or quantity of food.
Insight #3: Going too crazy with OMAD food/carbs, especially eating too late, is very problematic for sleep and next-day workout performance
For example, here is a day where I went too crazy with brats and chips for dinner:
As an aside here, it’s interesting that eating peanuts doesn’t spike my blood sugar, but does seem to elicit a significant insulin response to significantly lower my blood sugar levels.
Anyway, eating too many refined carbs here (buns & tortilla chips w/ 12oz of beer) produced one of the most dramatic spikes in my whole 30-day experiment, which led to me getting super sleepy and falling asleep at the peak.
Interestingly, upon waking up in the middle of the night my blood sugar dropped significantly (perhaps some melatonin cleared to allow insulin release?).
Here is the gist of the sleep data from the above night (from the SleepWatch app on Apple Watch. As you can see, not great):
And workout data from the next day was similar to that shown above (i.e. low energy, struggled, etc..)
Insight #4: The sweet spot for me on OMAD seems to be a reasonably early dinner with plenty of carbs (but not too much!)
Here is a typical “good” OMAD day that leads to decent sleep and a solid workout the next day:
For dinner I had veggie straws as a snack, turkey sausage burritos with avocado, low-carb tortillas, corn, and cheddar cheese; plus peanut M&Ms for dessert to satiety (ended up being two small bowls full).
Here’s what sleep data looked like. I intentionally didn’t pick out a day with over 8hr of total sleep, as this is more closer to a typical “good” and realistic night:
^^ Oura ring for overall night and resting heart rate.
^^ And this kind of profile for me for HRV is common during the course of the night (of course, everyone is different).
Also, when the G6 said 118 that morning, my manual meters said 100 and 99 (hence the recalibration).
And then the workout (trail run) that day felt great:
Additional miscellaneous insights
At various times throughout the experiment I tried eating breakfast, lunch, moving the time window later or earlier, trying different foods, etc… and what I learned was the following:
- Eating earlier in the day results in much smoother glucose rises and falls.
- Hot showers raise blood glucose (and I learned in summer 2019 that cold showers lower it)
- Eating both lunch and dinner (or both breakfast and dinner) doesn’t necessarily change the volume/mass of food that my body naturally wants to eat at dinner.
- I know this sounds crazy, but I can’t imagine eating three meals a day now. That’s a lot of food (and time).
- Overall, “low-carb” foods do seem to work as advertised. I was skeptical about low-carb tortillas, for example, but they do seem to not spike blood sugar as much as normal tortillas.
- However, as I mentioned above, if I don’t get enough carbs or mass at dinner then I get a bizarre empty/low-fuel feeling.
Overall, the relative randomness and variability of Whoop and Oura’s data integrity make me much less of a fan of them vs. the Apple Watch. I’m excited to check out future versions of Whoop and Oura, but for now what the Apple Watch offers is sufficient for my ongoing miscellaneous N=1 experiments.
Also, what I learned regarding optimizing sleep and workout performance on OMAD was a bit counter-intuitive. I had thought that limiting carbs and sticking to more “balanced” macros would produce better sleep and workouts, but that wasn’t the case. This is ultimately good news for me, as I enjoy eating a large dinner mostly ad libitum.
In a future study I want to look specifically at the effects of alcohol on sleep. I learned a long time ago that more than two drinks in one evening was problematic for me, but I’m curious if there is a way to isolate enough variables - and get reliable enough sleep-stage data - to learn where my personal limit is in terms of amount and timing of alcohol in the evening.
Author’s note: I recorded 144 pages of notes during this 30-day experiment, so there is a LOT more data I can (and will) likely extract and write about. Subscribe to my newsletter to stay tuned, and/or email me at email@example.com and I’d be happy to answer any questions. Thanks for reading!