WOW have you guys read the PCMAG review of nod32?

Now I would completely agree with that assessment. It's apparent from their own results that NOD32 has the best detection capabilities. But maybe not the best interface. And if your the average non-techie that reads the mag, PC-cillin might be the better choice. And in my opinion, they should select the best fit for their audience.

 

I also found that test result very unusual? I scan my HD periodically with AH and don't get FPs. Nor have I seen that high FP problem posted here? Perhaps someone that is seeing that high number of FP can post a screen capture.

 

tempnexus -

can you document that infection? So far we have not been able to actually get definitive information on exactly which (ITW) virus NOD32 does not catch. People post something to this effect every now and then but upon asking for the name of the virus no one seems to be able to provide the information for further research.

Dazed and Confused -

In testing there are zoo virii which are created for testing purposes and are not the same as In The Wild virii. Zoo virii are created to challenge an AV program in a lab for testing purposes and these will not be found floating around in the real world. ITW virii on the other hand are real threats that infect computers and are found floating around in the real world. These ITW virii are captured and so far no one has yet to document when any ITW virii has been not stopped by NOD32. NOD32 stops real threats and not necessarily zoo (lab) virii because zoo virii are not a threat to someone's computer unless that someone takes their computer into the AV testing lab and exposes their computer to the zoo virii.

Some people test AV programs using a combination of zoo and ITW virii and then claim because some AV program did not stop certain zoo virii (but stopping all ITW virii (real world threats) that the AV program is not very good as some other program that has stopped more zoo virii (but not as many ITW virii).

Virus Bulletin in awarding NOD32/ESET with more 100% awards than anyone else indicates via their independent (third-party) testing that NOD32 stops more ITW virii (real threats) than any other AV program that has been tested by them. Virus Bulletin publishes their testing methods for all to see. There are no independent (third-party) testers who publish their testing methods and who test for real-world virii that find NOD32 is surpassed by any other AV program in real-world Anti-Virus protection. If this is not so we invite anyone to document your claim so we will stand corrected.

BTW - Virus Bulletin has some information about their testing methods here. The 100% award is given only if the AV program has detected all ITW virii during both on-demand and on-access scanning and generate no false positives when scanning a clean set of files.

Lastly - PCMag and PC World are two separate magazines. The one with the article in question is PC World.

Best wishes and may everyone reading this be virus-free!

Be seeing you

 

Q - Thanks for clearing up this up for me. I was wondering what the difference was. Learning more from this forum every day!

 

Yeah, Stan. I should have said "extraordinarily high" or "outrageously high." That many fp's just looks bizarre to me. Especially without a correspondingly publicly known reputation for a high rate of fp's. That's a real head scratcher for me.

So just to clarify, while some of us are discussing the PCWorld June 2004 article that was linked to here, it appears that the original post by Tempusnexus refers to a PCMag June 2004 article that apparently is not available online, correct?

 

The June 2004 issue of PCMag does not seem to have an AV test in it.

Be seeing you

 

To those who believe ANY test by AV-Test.org GmbH is or has ever been accurate, credible, or valid in any way, shape, or form, all I can say is "ROFL".

 

If you give the slightest credence to tests by Andreas Marx, it is you who should try living in the real world, my friend. His past history of poor quality tests is a matter of public record. Leopards do not change their spots.

 

It seems there is indeed bad blood between Andreas Marx and NOD32. Be that as it may, I take issue not with his treatment of any specific program, but with the poor quality of his testing.

Summoning a demon is seldom wise.

 

Unless methodologies, authenticated samples, and control agents are made available for independent peer examination and review, no test can possibly be accepted as valid by academia or the scientific community. I am quite surprised to see Magdeburg lending its name to those tests.

 

Who won or lost is not the issue, dear lady. The issue is the validity of the tests themselves. As many here have pointed out, and no doubt many more will, Andreas Marx has the un-enviable history of poor quality testing.

 

And to that I will add my own "ROFL".

 

I think that VB means more than this test, personally. I agree with Paul's assessment of things. They can give the the results and then post a link to more info on the test - for people who want to see it.

I just bought NOD32 last night, so I'll be stopping by more. KAV and NOD32 on a system is a great combo. Now I have to decide which to use for on-access and which for on-demand. Was not terribly impressed with KAV 5 and went back to KAV 4.5.

 

Hi there,

in what kind of test protocols are you interested in? I think, they have EVERYTHING publicly available at their website!

A test description (a German version is available as well):
http://www.av-test.org/down/data/2000-02-en/description.txt

Even better: Lots of papers, about dialers, test lab installations, test strategies, false positives and so on. At least 100 pages to read!
http://www.av-test.org/sites/references_papers.php3?lang=en

Why do you think, that's not enough?

Wanadu

 

That first link refers to a test in 2000. It does not provide specific info about this test.

The rest is archived material, articles etc.

Where is the specific info about the methodology, viruses and clean files tested, etc. for the test cited in PC World? If it's not available, the other material is certainly not enough and not specifically relevant.

 

Did the PC World article state their test was a benchmark for future AV testing? No. Did they say it was the most extensive and exhaustive AV test ever conducted? No. Did they specifically say it was done in a scientific way? No.

IT WAS JUST A TEST !!

If I took a ITW virus at random and introduced it to each one of my 3 PC's, each running a different AV, in my opinion that would be a VALID test. Would I call it scientific? No. But the results would definately influence what AV I'm going to use in the future. Should the results of this hypothetical and VERY limited test pursuade someone else to use or not to use one of the AV's tested? That would be for you to decide. But don't call my test biased!

At the other end of the spectrum, if they used every ITW virus, and tested them on every type of PC ever made, and used every accepted scientific method ever published, the results of that test would NOT, in my opinion, be as valid for my system as the hypothetical one I conducted above.

Hye, I'm a NOD32 user too. Just get over it!

 

The article as I understand it at least has little to do with "ITW" viruses. The minimum criteria for the apps PCWorld featured in its article was they all detected the ITW viruses as noted previously. Zoo viruses and trojans presumably are what are primarily featured in the PCWorld article if I'm correct in my reading. (In addition to testing for heuristics and false positives which resulted in curious results in my view.) One of the problems with tests is that if the perameters of the test and methodology is not made clear many casual readers don't necessarily understand what the test results really indicate and so go away with a completely wrong impression or inaccurate view of products based on a test they've seen.

Furthermore, PCWorld didn't conduct the test: AV-Test did. Since they presume to be reputable testers striving for an international reputation and hire out their services for such purposes, such questions regarding their testing methodology and credibility are important IMO. So yes, in such cases it does matter if the tests are performed scientifically as opposed to what any Joe Schmoe can do on his own, reporting anectdotal results. it is also important if others are using such tests to either decide what products to use for themselves or to recommend for use.

I don't lose any money off these tests as some AV vendors on occasion do. Their concerns are understandable if some well known public venue publishes such tests and it turns out they were crap or inherently flawed. In that case any concerns about the validity and credibility of the tests are literally a bread and butter issue.

But some shoddy testing is used by some people to decide what AV to use or recommend. So I think it does matter in more than just an academic sense. Especially when the test is performed by a testing organization such as PCWorld used in order to give the appearance of some measure of credibility and reliability.

 

How inaccurate a view of products could they get because of this test? If they bother to read the results, they would see that NOD32 has the best scanner of all tested. Do you disagree with that conclusion?

 

But others base their decisions on it. Such a test can have a pretty large impact. The problem with the Andreas Marx Tests is, that you get different results - it only depends for which magazine he wrote (made) the test. Sometimes i wonder if he realy understood the programm and knows how to use it. Another problem is, that many magazines introduce him as "the leading AV-Expert" and the average user believes it. I´m not a VB or NOD fan at all, but even if he would rank my favorite AV as number one - i wont trust his tests. Too many questions left unanswered. For example - the false alerts - to me it smells like a specially prepared test-set - but we will never know, cause all this important information was not published. To me - and i realy don´t want to offend anyone - he is just a ghostwriter with a "extended" knowledge about AV-related products.

It´s not only a matter for NOD32 - imho it´s dangerous if so many reviews in different magazines are made by a single person in such a way.

 

Frankly, I don't see how you arrive at that conclusion based on the chart that shows higher detection rates in the various categories for both Norton and McAfee with considerably fewer false positives. Is there something I'm missing on this chart? http://www.pcworld.com/reviews/article/0,aid,115939,pg,5,00.asp

 

I'm combining detection performance and scanning speed. Both better than their top pick.

 

If this happens, can it really be that bad?? From what I've read in the paper, lots of folks don't even have an AV. It's quite possible if you conducted the test more scientifically, the results would be different a few percentage points or so. Maybe the test is not perfect, but for the average newbie that reads that stuff, it's probably not bad advice.

 

Well - if you see it like that then there isn´t a test needed - The words USE A VIRUS SCANNER - IT´S IMPORTANT should do the trick