October 2020 I got my Indiegogo-backed Amazfit X, then in December I got a Amazon Halo band, and they both seem like pretty cool gadgets. There is some overlap in what they measure, including both steps and sleep quality.
In sleep they both provide an overall sleep score and how long you slept, and then sub-details such as the time spent in each phase of sleep i.e. light sleep, REM, and deep sleep. The data and information is supplied numerically and graphically. To check out how close the two sets of measurements were I compared the daily records for one week.
Right from the first day there were indications of a problem, the two sets of numbers diverged widely! For example for the first night:
- The Halo said that I had slept for 4 hours 48 minutes, the Amazfit X for 7 hours 44 minutes
- Halo had 3 hours 40 minutes (63%) of light sleep; Amazfit X reported 3 hours 50 minutes, or 49%
- Halo had 1 hour eight minutes (24%) of REM; Amazfit X had 2 hours four minutes (27%)
- The Halo showed zero deep sleep; the Amazfit X had 1 hour 50m minutes (24%)
The differences were as startling every night of the seven days that I compared! So now I found myself in a quandary and unsure of my sleep numbers - was the Halo correct, was the Amazfit X correct, or were both of them perhaps wrong?!
In the end I decided that it was the Halo that had to be incorrect mostly because it claimed that I only slept below 5 hours a night for the entire week while the Amazfit X reported that I slept between seven and eight hours every night.
Given that I generally went to bed between 10 and 10:30 pm while getting up at around 7 a.m. (i.e. a period of eight and a half hours), for the Halo to be right I'd have to be awake over three and a half hours every night. And while it is true that I did get up several times (interestingly, here they agreed I was up 6 times) every night, each was for a very few minutes and I always fell back asleep in a very few minutes!
More importantly, the deep sleep time reported by the Halo TOTALED one hour and 20 minutes for the ENTIRE WEEK. I looked it up and apparently "The average health adult gets roughly one to 2 hours of deep sleep per 8 hours of nightly sleep." The AmazfitX showed a few days where my deep sleep was on the low side (15%), and others where it was higher (up to 30%), which seemed much more likely!
Next, while the step counts were closer, they still occasionally showed quite different numbers, for example on this date they diverged by a significant 20%!
The Halo also can measure body fat percentage; you take photos of yourself and it uses an algorithm to calculate body fat. I tried this out, and was appalled by the result. Now I'm not in particularly good shape but at 5 foot nine and 160 pounds I am in the 'normal' range for BMI... True, it's closer to the top of the normal range (18.5% o 25%), but still. My scale reports my body fat percentage as 19.5%, however, the Halo estimate came back at 31%... more than 55% higher!
The end result? The Amazon Halo has been relegated to the "closet of misfit gadgets!"