Accuracy of Apple Watch Vs Amazon Halo Step Count

We make a big deal out of finding the accuracy of the metrics that fitness trackers are monitoring. It seems only right if you are going to take some action, or allow a tracker to dictate your activity, that the outputs it provides are both consistent and reliable. For metrics like VO2 max accuracy this means ensuring that measurements makes sense and, since they are approximates, relate well to what you would see in a true lab test. But for simple metrics like step tracking this is even more critical. 

While many trackers are pretty consistent day to day, there is also concern that different devices get different readings. Maybe you are taking part in a workplace fitness challenge for total step count, or you are just using a daily tracker to encourage you to beat a single friend. No one likes to find out that they may be losing their challenge just due to the device they picked. We reviewed Fitbit step accuracy separately, but also had access to both and Apple Watch and Amazon Halo band. We used these two to wear on the same day, doing various activities, and find out how they compare to each other in terms of their total step counts.

We measured this accuracy under the following conditions. The Apple Watch worn on non-dominant (left) hand, and the Amazon Halo on the dominant (right) hand and counted as such in the app. Each was worn for the entirety of the day, put on 5 minutes after waking and taken off 5 minutes before bed. Devices were both charged overnight. The results for Apple Watch step accuracy vs. Amazon Halo accuracy:

• Short Day Accuracy: Apple Watch 2730 steps, Amazon Halo 2964, Difference 8%
• Medium Day Accuracy: Apple Watch 11429 steps, Amazon Halo 14296, Difference 25%
• Long Day Accuracy: Apple Watch 26925 steps, Amazon Halo 30160, Difference 12%

It is worth noting that “accuracy” is not an ideal term here. While we have compared individual activities (a 20 minute run while counting steps) and found the Apple Watch to be closer to the count, we did not have a known count for each of the days. We assume, based on the other observations, that Apple Watch step counter accuracy worn on a non-dominant hand to be the closest to a true count. 

Apple Watch Step Counter Accuracy vs Amazon Halo On 3 Days of Testing

They are different devices so it comes as no surprise that the Apple Watch and Amazon Halo band did not report the same steps. On a short day of steps, without a run, working at a desk, and mostly being sedentary, the Amazon Halo tracked 8% more steps than the Apple Watch. This isn’t too much variability between the two, and might be explained in part because the two were worn on different hands. The Halo was on a dominant hand (used for writing and mouse clicks). This was an abnormally inactive day, just walking around the house and doing typical tasks, while mostly working at a desk.

The accuracy between the two devices got worse as the activity increased and became more vigorous. On a day with a workout, in this case a single 5 mile run mid day. Here the two devices were much different, with the Amazon Halo band measuring 25% more steps than the Apple Watch. This is the largest difference observed ever while wearing the two devices simultaneously. 

At least in a small sample the accuracy did not linearly increased, so there is some noise to this. The final day compared here is a big one. The Halo band measured 12% more than Apple Watch, on a total of somewhere between 25-30k steps. The day included two runs and a long walk, or a total of around 14 miles of activity in addition to steps taken meandering about the house. 

Step trackers are great, however this simple experience shows an important point. The absolute step count may not be as important. In either case the goal is to track, mostly, to encourage staying active. Regardless if the step tracker is getting you from 8000 up to 10,000 steps that’s great!

It doesn’t matter if you don’t actually take 10k steps. Even if it’s really only moving from 6k steps to 8k steps, but telling you it’s 10k, that’s still better than 6k and not knowing! 

What effects step count accuracy?

It’s no surprise that step trackers don’t return the same number every time. Each of these works on an algorithm that finds how much the device moves, using an accelerometer sensor. That signal from the sensor then has to be trimmed for noise. Raising your hand over you head, swinging an arm during a run, or moving your hand while grilling a mouse all cause some signal to the step tracker. Sometimes the signal might not get counted right.

Step counts should be expected to vary based on the hand they are worn on. They could be effected by the type of walking you do (up or down hills, hiking over non steady surfaces) too. 

In this simple experiment it was encouraging to see that although step trackers may not be accurate, they are precise. Precision should imply that the step tracker is accurate, when compared to itself, even if it’s not a 100% representation of a “step count”. In this case you can think of step trackers just as activity trackers that happen to correlate nicely to the number of steps you took. In part this is just naming conventions. 

As we get more ways to track steps we have seen similar results. Ultimately, the accuracy of the individual step tracker is not the one feature to rely on. The other features of an activity tracker, it’s app interface, and the form and style, may be much more important in deciding which device to make your daily choice.

Apple Watch Vs. Amazon Halo Accuracy in Other Metrics

Hopefully this experiment encourages you to keep active and make good decisions based on the metrics you know. The accuracy of devices is not everything. That said, we have also looked at these devices for other metrics as well. You can check out the accuracy of various body scan techniques including the Amazon Halo body fat accuracy. You can also check out how these two devices stack up when measuring sleep patterns.