A method for estimating the hazard rate, or probability distribution of a random variable observed subject to data truncation. It is closely related to the product limit estimator for censored data.