Motion Algorithm

Motion Score / Activity points

Motion’s proprietary activity points system can be summarised into these key steps:

1. Gather data from tracking sources

   • Motion collects user’s health and fitness data from a myriad of sources configured by the user.

2. Clean and normalise data

   • The input data is cleaned, and normalised into a consistent format for further processing. Here, data may be adjusted for motion to account for tracker biases, based on a model trained on millions of workouts tracked on the motion platform.

3. Resolve data conflicts

   • Since user’s often have more than one tracking source, duplicate workouts / steps / etc, need to be systematically detected and combined or discarded.

4. Generate activity points

   • Now we have a clean representation of the user’s data, it can be converted to Activity Points. A number of data transformations happen to assign point values for activity. This takes into account, the user’s history within different activity types, general fitness levels, and biological markers.

Dynamic goals

Motion runs on weekly cycles, and each week every user gets assigned a new activity points goal.

A user’s score can move up, or down, depending on how they have been performing recently. Additionally, each user has a “safe-zone”, if they score within this zone their goal will not change. This allows them maintain a consistent goal if they are in a maintenance phase.