Some parametrizations of `NegativeBinomial` can underflow `p`

[theogf]

With the NBParamII for sufficiently large values of f (something like 800), p will be equal to zero and many estimator starts to break down. This might be something to be dealt in Distributions.jl though.