assumes that the model that the binomial distribution in each
year is the same is true. A comparison of the score and adjusted
prediction intervals reveals that the theoretical comparison of
Kullback–Leibler distances for the two predictive distributions
is consistent with the comparison from this application example.