A growing body of evidence has shown that incorporating behavioral economics principles into the design of financial incentive programs helps improve their cost-effectiveness, promote individuals' short-term engagement, and increase compliance in health behavior interventions. Yet, their effects on long-term engagement have not been fully examined. In study designs where repeated administration of incentives is required to ensure the regularity of behaviors, the effectiveness of subsequent incentives may decrease as a result of the law of diminishing marginal utility. In this paper, we introduce random-loss incentive -- a new financial incentive based on loss aversion and unpredictability principles -- to address the problem of individuals' growing insensitivity to repeated interventions over time. We evaluate the new incentive design by conducting a randomized controlled trial to measure the influences of random losses on participants' dietary self-tracking and self-reporting compliance using a mobile web application called Eat & Tell. The results show that random losses are significantly more effective than fixed losses in encouraging long-term engagement.