Arun Kumar Saha <arun.kumar.saha <at> gmail.com> writes:
Can the Statisticians here develop some good Statistical tool to stop this
Spamming keeping Type-I error almost zero?
As Bruce Schneier (surely among others) has pointed out, the difference
between security against natural risks (analogous to most statistical
problems) and security against attackers is that the attackers adapt
to avoid countermeasures. In the terminology of Hurlbert (1984), a classic
in ecological statistics, this is the difference between "non-demonic"
and "demonic" interventions; the tools of statistics are generally intended
to guard against non-demonic intervention only.
Ben Bolker