Self-consistency of Kaplan–Meier estimation

The Kaplan–Meier estimator is self-consistent under fractional allocation of censored event times according to its own conditional fitted event distribution. Adding the corresponding fractional event counts and risk set weights leaves the estimator unchanged.