Reply to "Rebuttal to Replies I and II by Varotsos et al." by F. Mulargia, W. Marzocchi and P. Gasperini

Mulargia et al. [1996] claim that earthquakes (EQs) can be “predicted” (in retrospective) “much more efficiently than VAN” using a “rule”, they obtained from PDE catalogue. We show that this claim is undoubtedly wrong. Their “rule” issues a great number of false alarms, which exceeds that of the “predicted” EQs (mainly aftershocks) by a factor larger than 10. The errors diagram recommended by Keilis‐Borok [1996], reveals that Mulargia et al.'s [1996] “rule” corresponds to a non‐meaningful algorithm; on the other hand, this diagram reflects that VAN is meaningful.