Released yesterday, the above URL will render the PDF:
"Malware Removal Test"
Unfortunately, and the report points this out also, only /ten/ (10)
different types of malware were selected for the test. Maybe if the
numbers of malware had been increased ten fold (or more) the results
might have been different.
Because of the small statistical sampling, I believe the report is
somewhat flawed and should only be used as a general guide.
Of course the test failed to test some of our favorites around here.
Pity.