One of the key features of the Apple Watch is its ability to serve as a fitness tracker, replacing devices made by Fitbit, Nike, Jawbone, and others. Fitness trackers are often quite inaccurate; I’ve tested several, and only the Fitbit One counts steps accurately. Some fitness trackers also have heart rate sensors, as the Apple Watch does, but I’ve never tested any of those.
But how about the Apple Watch? Is it accurate?
I’ve had my Apple Watch for three days, and I’ve been recording my activity, and comparing it to the Fitbit One. I’ve found that, in some ways, the Apple Watch is very accurate; in others, it’s all over the place.
The Apple Watch records different types of metrics, not just steps. While it does keep a step count, it focuses on three metrics to determine your activity. You can see them in the three rings that display on the Apple Watch, and in the Activity app on an iPhone:
The outer ring shows the number of active calories you have expended, and the ring is based on a goal you set (I set mine to 500 calories, and increased it after the Watch prompted me to on Sunday evening). The second ring is active minutes; it is measured against a goal of 30 minutes a day. The third ring is standing time: it measures whether you have stood for at least one minute in each of twelve hours of the day. As you can see, all my rings were will beyond their goals yesterday. When you reach a goal, the end of the ring is at 12 o’clock; as you exceed your goals, the ring keeps turning.
The Activity app also shows you more detail: the number of calories, number of active minutes, workouts (if you’ve used the Workout app on the Apple Watch), and the number of steps you’ve taken, and their distance. It’s this final metric that allows one to compare the Apple Watch with other devices.
Since I’ve had the Apple Watch, I’ve also been wearing my Fitbit One. On Saturday, the Apple Watch recorded 8,300 steps, and the Fitbit counted 8,480. On Sunday, the Apple Watch counted 7,938 steps, and the Fitbit 8,409. In my tests, fitness trackers that you wear on your wrist are very inaccurate, and all the devices I tested over-counted steps. The Apple Watch, however, is under-counting. The Saturday number was very close, but Sunday’s number was about 5% less. I suspect that the Apple Watch is comparing movement data with what is recorded by the iPhone’s motion co-processor, allowing it to be more accurate than standard wrist-worn devices, but it will never be as accurate as counting steps as a pedometer that you wear on your belt (such as the Fitbit One).
But Apple doesn’t use steps as the main metric; I think the only reason they show the step count is because people are used to seeing this number with a fitness tracker (with the notable exception of the defunct Nike+ Fuelband). Apple is focusing on calories, and that’s the main goal you set.
Apple and Fitbit clearly use different calculations for calories. The Apple Watch has the advantage of being able to check your heart rate (the only Fitbit device that can do this is the Fitbit Surge). It checks your heart rate every 10 minutes, or more or less continuously when you perform a workout using the Workout app on the watch (unless you’ve put the Workout app in Power Saving Mode, in the settings). But how accurate is the Apple Watch’s heart rate sensor? For the most part, when I check it, the numbers look good. But, at times, it’s way off.
When walking on my treadmill yesterday, I checked my heart rate in the Workout glance on my Apple Watch. At times, it was around 100, which seemed to be correct. But several times, it was very low, such as 64 or 72. At one point, I took my heart rate with Withings’ HealthMate app, which lets you measure your pulse by placing your finger over the camera lens of the iPhone. At the left, the Apple Watch; at the right, the Withings HealthMate app:
Also, one time when I was walking outdoors, I checked my heart rate on the Apple Watch, and it said 153. I took my pulse just in case, but it wasn’t that high, it was around 110. The Apple Watch got the number right a couple of minutes later.
With this in mind, I looked at the calorie calculations for both the Apple Watch and the Fitbit. On Saturday, the Apple Watch credited me with 687 active calories and 3,180 resting calories, for a total of 3,867. The Fitbit app calculated that I burned 2,742 calories. On Sunday, the Apple Watch counted a total of 3,915 calories, compared to 2,752 for the Fitbit app.
I’m not sure how these numbers can be so widely divergent. I tend to think the Fitbit app is more realistic (neither can be accurate, since measuring calories is a bit of black magic). Calorie calculators for my age, sex, and weight, show numbers closer to those of Fitbit.
Counting Active Minutes
Another metric that the Apple Watch counts is active minutes. As Apple says, the Exercise ring:
“displays how many minutes of brisk activity you’ve completed towards a goal of 30 minutes. Every minute of movement that equals or exceeds a brisk walk, whether it’s working out or playing with your kids, counts toward your Exercise goal.”
As such, any brisk walking should count as active minutes. In my tests, I found that I could take a walk at a steady pace, and some of its minutes would count, others not. For example, yesterday evening, I walked outside for 21 minutes, with my heart rate around 105. I was warm, nearly sweating, so I think this counts as brisk. The first 5 minutes were clocked as active, but the rest of the walk wasn’t. If I walk on my treadmill, some of the minutes count, and most don’t.
Apple says that you should calibrate your Apple Watch, for better accuracy. This isn’t complicated: just walk or run outdoors with your iPhone for 20 minutes or more, and the Apple Watch should calibrate using the GPS to determine how long your stride is at different speeds. I’ve done this, twice.
A number of users have chimed in on a long thread on Apple’s support forum. Some people suggest that active minutes depend on your heart rate, that it should be in the aerobic level; that’s not what Apple says. Brisk activity is not aerobic. It’s clear that some activities will pose problems to the Apple Watch. Since it can use GPS to track you, it can determine your pace, and how much distance you cover, when you’re walking or running outdoors. But for indoor activity, such as a treadmill, it can only use the speed of your steps and your heart rate. So, while it will calculate calories for such activities, it won’t calculate active minutes. (Though you can override this with the Workout app by recording your activity as an Other workout. All its minutes will count.)
In addition, I’ve been watching when it records active minutes. Yesterday, for example, it recorded a few active minutes in a supermarket. I wasn’t excited enough about the tomatoes I bought – even though they were very good – to make my heart race for several minutes, yet the Apple Watch counted some of that time as active. Just as it did when I was watering the lawn on Saturday.
The Fitbit records active minutes, presumably just based on pace, since it has no other sensors. On Saturday, it recorded 47 active minutes, and on Sunday, 33. The Apple Watch recorded 31 minutes on Saturday, and 50 on Sunday, but that’s with my using the Workouts app to “force” it to record my minutes on the treadmill, so I can’t really compare which is more accurate.
In a walk I took today on my treadmill, using the Indoor Walk workout, for thirty minutes of activity, only 5 were counted as active. Here’s what my Exercise ring looks like after the walk (I don’t know what I was doing earlier today to get one minute of activity):
Since I was walking at a constant pace, with a heart rate that was roughly constant after the first couple of minutes, why didn’t more than six minutes count as exercise? Could it be that the heart rate sensor, which clearly has some issues, meant that it was only accurately recording my heart rate for a few minutes?
If I look at the Health app, it shows several readings per minute for my heart rate. If you look here at 12:44, you can see that it’s clearly not getting it right. It’s worth noting that I’m wearing the watch snugly, so much so that I can’t fit a finger between my band and my wrist, but it’s not too tight either.
There’s also another, longer period, from 12:52 to 12:54, where my recorded heart rate dropped from 100 to 87, then, at 12:55, it went up to 108. It stayed from around 103 to 108 for the remainder of the walk, which ended at 1:00.
The Apple Watch also seems to have difficulty recording distance and pace for an indoor workout. This shouldn’t be hard. The Apple Watch knows the approximate length of my stride, especially since I calibrated it outdoors with GPS. So it should calculate distance and pace based on the number of steps I take, multiplied by my stride. In a half-hour walk today, it recorded a distance of 1.5 km, whereas my treadmill recorded 1.7 miles (about 2.7 km). While I was walking, I checked the pace, and it was all over the map, ranging from 20 min/km to just over 9 min/km. I was walking at 3.6 mph, so the actual pace should have been around 10:20 per km. Sometimes the Apple Watch lost track of my pace entirely. I took a number of screenshots while walking; they’re in time order, as you can see in the upper-right corner of each image. (I started my workout around 12:30, and only started looking at the Workouts app after about 20 minutes of walking.)
At the time of this writing, after I walked on my treadmill for a half hour, the Apple Watch shows that I have walked 1.58 miles, and the Health app, where I am able to change the units, shows a total of 2.56 km. The Fitbit app, which simply calculates steps multiplied by stride (based on a stride I measured), says I walked 3.22 km, which matches more correctly with what my treadmill showed. The step count is nearly the same: 3,524 on the Apple Watch, and 3,575 for the Fitbit.
I’ve said before that fitness trackers are more about motivation than accuracy. If they get you to be more active, by prodding you to reach new goals, then they are successful. They should give you reliable data about your activity, though, and not be far off the mark. (I consider that the 5% difference in step count is acceptable.) But they shouldn’t lead you to wonder whether your activity is counted correctly, as the Apple Watch does. Either the Apple Watch is severely flawed in its fitness tracking capabilities, or I received a dud (I’m going to call Apple later to try and find out). I’m curious as to whether other readers have compared the Apple Watch with other fitness trackers, or whether anyone has similar data, reported by the Apple Watch, which just seems wrong. Feel free to post in the comments if you do.