Example for how entropy is to be interpreted. In this mock example for the classification of two reviewers, similar to Table 1, all probabilities over configurations (first column) are equally likely and thus the entropy is ascribed its maximal value of 2.

The second column shows a case where the first reviewer is not yet classified, but is considerably more likely to belong to one class, hence entropy takes on some value between 1 and 2. The third column contains an example where the first reviewer is fully identified, but the probability does not favor any classification for the second reviewer leading to the entropy value of 1. The fourth column has an example where the first reviewer is fully identified, but the probability favors one classification (rival) for the second reviewer, leading to the entropy value between 0 and 1. The last column contains an example where one configuration has probability 1, hence the reviewers are fully classified, leading to 0 entropy.