Estimation problems for the two-parameter negative binomial distribution
This note shows a major problem in the proof of Anscombe's conjecture given in Aragon et al. (1992). It also proves that the unbiased estimator of the parameter k does not exist.