Acceleration and de-acceleration events are generated using a proprietary algorithm. The algorithm takes six different axes into consideration (along with other variables, like speed) to generate events. The algorithm uses data that Lytx has accumulated over 20 years of analyzing accelerometer data, including over 5 million events that occur every month. Using this algorithm, risks are determined with much greater accuracy and false positives are reduced accordingly.
A slight de-acceleration is less relevant at 10 mph, and more relevant at 70 mph, even if the g-force during both is equal. Therefore, the algorithm is not based on g-force. Events are generated through a combination of sensors and data streams that determine what risky driving looks like.