In previous posts, I’ve observed my inability to replicate Mann’s verification statistics, the source code for which was not archived despite representations to the contrary in the original article.
Mann has issued another uninformative correction notice (bearing the date Dec 1, 2008) which states:
UPDATE 1 December 2008: Supplementary Figure S8a had a small error due to improper calculation of the validation statistics. The corrected figure can be found here (PDF)
OK, now, pretty please, what are the properly calculated validation statistics? And, pretty please, where did the error occur?