07/06/01 14:05:29
>>435-436
URLﾘﾝｸ(www.wilderssecurity.com)
> Percentages are small, but you should remember that nowadays most computers
> are exposed to a vast amount of executable files/word documents/scripts/etc,
> and that most users have to rely on their AV decision. And these computers are
> exposed to much more legit files than to malwares.
> Most users will not encounter more than 10 different malwares per year.
> But they may use more than 200 different programs/documents/scripts in the same period.
> If an AV has an overall detection rate of 99%, and considering that the malware
> they face are taken at random from the pool of existing malwares (that assumption is clearly wrong),
> the probability they get infected is lower than 10%.
> Now, if the same AV has a false positive rate of 0.1%. Under the same assumption of "perfect sampling",
> the probability that they will hit at least one false positive is higher than 18%.
> Since most users tend to panic when they think they are infected by a virus (think to jdbmgr.exe),
> it is very important to minimize their occurence.
That's why I fully support the decision of IBK to consider it as an important parameter for evaluating the proactive detection of antiviruses.