On Wed, 1 Oct 2003 20:03:18 -0400, Kee Hinckley wrote:
At 12:21 PM -0500 10/1/03, Terry Sullivan wrote:
Data from multiple independent sources (including Liam Meany, Scott
Nelson, and myself) indicate that "otherwise identical" trap
addresses receive *vastly* different amounts of spam traffic.
[snip]
It's frequently been said that there are really only about 200 major
spammers. And certainly there are fewer types of spamming software
than that, and fewer sources of target addresses.
[snip]
...And it's definitely not a random population.
As you suggest, there are lots of possible/plausible explanations
that may actually account for the "unexplained" variance in _ad lib_
spam volume. And I absolutely agree with your core point, that said
variance is so obscenely large that it cannot possibly be truly
"random" in nature. Clearly, one or more systematic forces are at
play; but those systematic forces lie wholly outside experimental
control.
Still, it remains statistically "bad" variance, precisely because it
isn't subject to experimental control. By implication, any "live"
experiment simply *must* be able to factor it out at analysis time.
Otherwise, there's no point in running the experiment.
(All of which underscores the need for a trap-address effort.)
- Terry
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg