A New, Discrete Model of Lindley Families: Theory, Inference, and Real-World Reliability Analysis

Recent developments in discrete probability models play a crucial role in reliability and survival analysis when lifetimes are recorded as counts. Motivated by this need, we introduce the discrete ZLindley (DZL) distribution, a novel discretization of the continuous ZL law. Constructed using a survival-function approach, the DZL retains the analytical tractability of its continuous parent while simultaneously exhibiting a monotonically decreasing probability mass function and a strictly increasing hazard rate—properties that are rarely achieved together in existing discrete models. We derive key statistical properties of the proposed distribution, including moments, quantiles, order statistics, and reliability indices such as stress–strength reliability and the mean residual life. These results demonstrate the DZL’s flexibility in modeling skewness, over-dispersion, and heavy-tailed behavior. For statistical inference, we develop maximum likelihood and symmetric Bayesian estimation procedures under censored sampling schemes, supported by asymptotic approximations, bootstrap methods, and Markov chain Monte Carlo techniques. Monte Carlo simulation studies confirm the robustness and efficiency of the Bayesian estimators, particularly under informative prior specifications. The practical applicability of the DZL is illustrated using two real datasets: failure times (in hours) of 18 electronic systems and remission durations (in weeks) of 20 leukemia patients. In both cases, the DZL provides substantially better fits than nine established discrete distributions. By combining structural simplicity, inferential flexibility, and strong empirical performance, the DZL distribution advances discrete reliability theory and offers a versatile tool for contemporary statistical modeling.