The Gompertz demographic model describes rates of aging and age-independent mortality with the parameters alpha and A, respectively. Estimates of these parameters have traditionally been based on the assumption that mortality rates are constant over short to moderate time periods. This assumption is questionable even for very large samples assayed over short time intervals. In this article, we compare several methods for estimating the Gompertz parameters, including some that do not assume constant mortality rates. A maximum likelihood method that does not assume constant mortality rates is shown to be best, based on the bias and variance of the Gompertz parameter estimates. Moreover, we show how the Gompertz equation can then be used to predict mean longevity and the time of the nth percentile of mortality. Methods are also developed that assign confidence intervals to such estimates. In some cases, these statistics may be estimated accurately from only the early deaths of a large cohort, thus providing an opportunity to estimate longevity on long-lived organisms quickly.