Telematics data collected at the end of each trip is used to generate variables empirically proven to be predictive of future collisions. These variables (which represent the 5 key categories of driving behavior: Acceleration, Braking, Cornering, Distraction, and Speeding) contribute to the score’s calculation. These variables primarily focus on the recency and severity of negative events, such as the prevalence of harsh braking and evidence of excessive speeding, and consider positive events as well, such as the prevalence of smooth stops. Each characteristic’s value is then placed in different levels, and each level is assigned a unique score weight (coefficient). The coefficients are summed to essentially arrive at an overall score.