[PLUG-TALK] Prevalence of "owned" machines

[Keith Lofstrom]

On Fri, Jan 29, 2010 at 11:02:45AM -0800, Ronald Chmara wrote:
> 48% here:
> 
> http://indyposted.com/9937/nearly-half-our-computers-are-infected/
> 
> I suppose it depends on how "owned" is defined, as well.

Thanks for the pointer - it explains a lot.  The author of report is
a PR agency working for vendors of malware detectors.  They are good
at promoting their client's interests.

The Panda Labs "methodology" is self-serving and statistically invalid. 
Panda sells virus software, and their website offers a "click here for
a free virus scan" button.  They are actually reporting that half the
22 million clicks resulted in their software claiming to detect malware.  

Those clicks are nothing like a random sample.  That button is clicked
by Windoze users who (1) think their computer might be infected,
(2) probably don't already run software they are confident in, and
(3) run programs from random sites online.  This is a high risk
demographic.  People that get tested for bubonic plague are far more
likely to have the disease than the general population, too.

Rule of thumb: Unless they are measuring repeatable, identical events
with published methods, the more decimal places given, the higher the
bogosity.  This is true for all kinds of statements 96.3745% of the time.

All that said, thanks Ronald, for tracking down the likely source of
the number.  

I still would like to see some valid measurements, though it is hard
to imagine how a competent researcher would gather valid data.

Keith

-- 
Keith Lofstrom          keithl at keithl.com         Voice (503)-520-1993
KLIC --- Keith Lofstrom Integrated Circuits --- "Your Ideas in Silicon"
Design Contracting in Bipolar and CMOS - Analog, Digital, and Scan ICs