CORRECTION: In this paper (Hamilton, Shan, and Ziarko 1997), at the bottom of the fourth page and the top of the fifth page (pp. 333-334), there is a statement that: In Table 4, P('Y') = 7/12 since 7 out of 12 instances have Y as their decision value. Thus, without consulting the classification, one could estimate the probability of an instance having decision value Y as 7/12 = 0.6471. This should be amended as follows: In Table 4, P('Y') = 11/17 since 11 out of 17 instances have Y as their decision value. Thus, without consulting the classification, one could estimate the probability of an instance having decision value Y as 11/17 = 0.6471. I must have had a mental blackout when I put 7/12 instead of 11/17 because I used 11/17 in all my calculations (11/17 really is 0.6471 but 7/12 certainly isn't!)