The algorithms used to detect step tracking vary due to different factors. For wrist-based pedometers like the Moto 360, the algorithm may be affected by arm-swinging or the lack of it.
For example, if you are pushing a stroller or shopping cart, the lack of wrist movement will make it difficult for the device to track your steps. However, waist-based trackers like phones and traditional pedometers counter this by tracking steps using waist movement. Unfortunately, during certain activities like driving, where there is movement from the road, these trackers may miscalculate steps.
Wrist-based pedometers counter this by noting the lack of arm movement. These algorithms are merely a guideline for measuring health and physical activity. It cannot be taken as an accurate measurement.
While developing our device, we test it alongside foot-based trackers to ensure the most accurate results possible in various scenarios.
Our goal is to be within 10% of the actual step count in all cases. We also try to achieve even more accurate results for common tasks like walking outdoors.