Fig. 14

Wild boar (Sus scrofa), with scores being the probabilities; \(n =\) 1300; Kuiper’s statistic is \(0.07244 / \sigma = 6.688\), Kolmogorov’s and Smirnov’s is \(0.07149 / \sigma = 6.600\). The behavior in the present figure at the highest probabilities is similar to that in Fig. 13—the reliability diagrams with only 10 or 30 bins each underplay the very high deviation for the highest probabilities, smoothing away big deviation without warning. As in Fig. 13, the scalar summary statistics detect statistically significant deviation, while both are blind to the significant deviation for the highest scores that the plot of cumulative differences highlights; the steep drop at the highest scores in the cumulative plot has essentially no impact on the Kolmogorov-Smirnov or Kuiper metrics, unfortunately