This study investigates the impact of an incentive-based driving competition on risky driving behaviors using data from a large-scale naturalistic driving experiment. Leveraging smartphone telematics data collected from 116 drivers over 25,000 trips, the analysis focused on three key safety indicators: speeding, harsh braking, and harsh acceleration. A before-and-after design was employed, comparing performance in a baseline feedback phase versus a competitive phase featuring gamified rewards. Generalized Linear Mixed-Effects Models (GLMMs) were used to account for repeated measures and individual differences. The results reveal statistically significant reductions in speeding and harsh events during the competition phase, indicating that gamification can effectively influence short-term behavioral change. However, mobile phone use showed no improvement, highlighting the challenge of addressing distraction-related behaviors through competition alone. These findings support the use of dynamic incentives, such as those provided by the telematics app of OSeven, for promoting safer driving, while also underscoring the need for continuous, personalized interventions to ensure lasting behavioral improvements.