PC World AV Chart

I just paid for NOD32 a couple of weeks ago and I *did* jump out the window. Fortunately, the window was on the first floor and I was unharmed.

Still, I'm waiting for all hell to break loose. I'm scared to death that I'm going to get the Backdoor.Caligula-Hitler-Hussein-binLaden trojan that's circulating and, then, WHAM, my computers are useless and I won't be able to get any work done. After that, it's a slippery slope to criminality in order to make a living and I'll wind up in prison (not some white collar resort prison, either. I'll be going to Federal pound me in the a$$ prison).

 

Thats the same test, but the 84 you refer to is actually the overall rating of the software which was reviewed by PC World, not the malware detection rating which was performed by AV-test.

 

very interesting test

I wonder though, how do they test detection rates? Is it based on executing the samples or simply "dead" samples which are not executed? I think this would affect the results if they scanned "dead" samples, especially for something like KAV which includes a behaviour blocker. I can't really imagine them actually executing 1 million samples, lol.

 

I would be interested in AV-Test's testing methodology. For example, Avast and AVG scored 2/12 = 16.66% for polymorphic virus detection in the recent on demand AV-Comparatives test. AV-Test scored these two AVs in the upper 80% regarding polymorphic detection. That's a huge discrepancy.

 

I think that depends on the samples that was used on the tests...

 

As do the other numbers in the test.

 

Of course...

Why we should never see and trust in only one test?

 

I *think* that AV-test's polymorphic test set was larger than AV-comparatives. AV-comparatives uses a set of samples which are not in priority for many vendors, that is why so many don't do that well in detecting the polymorphics.

There have been far more polymorphics released than what you see on AV-comparatives, and it is possible AV-test included some of the other polymorphic malware as well. I'll see if I can get any further information.

 

Can you email Marx to send us LOG FILES of tested AV scanners?

I want to see 500000 in LOG FILES

 

Is there a spot on the AV-Test.org website to review test results for particullar Av's such as F-Prot?

 

I got a reply from Erik Larkin at PC World. As confirmed earlier, all products were tested at maximum settings for determining the detection rate (of course ). He also sent me the number of samples for each malware type that was used in the test:

File viruses: 191029
Macro viruses: 73117
Script viruses: 1184
Polymorphic viruses: 1895
ActiveX controls: 9
Backdoors: 107106
Bots / Zombies: 97215
Trojan Horses: 341993
Adware samples: 14517
Dialer: 56099

Total: 884164

Thanks to Mr.Larkin for the above information.
This test has been very interesting for me, lots of good surprises here and there.

 

For sure if you are talking about PC World. But the test in this thread was actually made by Av-test.org, so...

Maybe you are a bit right too, CAT QuickHeal won the latest VB 100 % Award 04-2007 but NOD32 did not, so the better av in this case was...

Best regards,
Firefighter!

 

The FP occurred just because of wrong heuristics settings used in the Linux version. The highest heuristic level was removed from Windows version about 2 years ago but for some reason it remained in the Linux version. Otherwise we would have gained the award again

 

The question to be asked here is WHY did this heuristic setting remain? Bad testing? Negligence? Low priority?

What I'm saying is that such mistakes ultimately cost Eset a VB100 award. Going to blame this thing and that thing isn't going to help (Not just Eset, but for anyone else), because in the end it was an Eset error anyway that caused the wrong heuristic settings to remain.

So, the fact remains, Eset lost a VB100. But yes, it still detected all ITW malware, and of course, this doesn't mean Eset or NOD32 sucks.

As for Firefighter's comment, you know he was joking Marcos, and all of us know NOD32 = 10000 times better than QuickHeal.

 

I didn't say it wasn't our mistake that that setting remained unchanged, I merely wanted to explain that it may not be a big thing if an AV misses a VB100 award except for marketing purposes.

 

Then I misinterpreted it, and thats my mistake. Sorry.

 

The number of adware and ActiveX samples seem low

 

Well i tested it with the eicar test and it did only detect it when it was om my harddrive! So no detection in the HTTP stream

 

After taking a closer insight, the REAL winner in detection was BitDefender, because it has the best OVERALL detection among categories. The LOWEST score in detection of categories with BDF was 95 %, be aware!

Best regards,
Firefighter!

 

Yes a good product if they could only get their resource usage down.

 

I have not tried the latest version of BDF, but how did Panda managed to get their product even lighter than NOD32? Before it was said to be a REAL resource hog.

Best regards,
Firefighter!

Btw, it seems more or less like a stock market, you have to get a new av every quarter to get the best profit!

 

Have been using BD10 AV for a few months now and I like it a lot. I run it on two systems, one with 1 Gig and the other with 256 Mb RAM and I find that BD runs light with noslowdown on XP.

 

Panda removed TruPrevent with Panda Antivirus 2007. And this is 85% of the reason why Panda is light now compared to before.

As for BitDefender, yes they need to fix compatibility problems and resource usage. Apart from that, if they continue on this level of malware detection, they are going to seriously make some heads turn.

 

I just performed the same test and it did give warning and blocked it before I answered the "do you want to run or save this file?" dialog box.

http://img255.imageshack.us/img255/1004/nortonwarnkw4.th.jpg

 

Ok, strange i can't remeber that it detect that when i tested it