AV Comparatives File Detection test for September 2014

And yet everyone instantly forget about the real-world tests where it had ZERO fp's...

 

When discussion is about wrongly classified files in "file detection" tests - why talk about FPs for websites? But if we talk about AV-C real-world test for September, I see a "1" for avast and not zero. But nevertheless here it has no bad FP score

 

Then it was one month back, i know it had zero once. But zero or 1 and 120 is a massive difefrence so you get the picture either way...

 

Well, AVG too has zero FPs in the real-world protection test, but I know that there are certain types of files for which AVG is quite trigger happy....

 

Well i always removed AVG because of moronic false positives on 3rd party stuff for games like compatibility patches for game Red Faction. Comodo never fixed that crap and neither did AVG. God forbid you actually download No-Cd patches because you hate using CD for your original games, because the mentioned pair will pretty much always detect it where avast! hardly ever or in my case, never. AVIRA gotten a bit better though still not that great...

 

I found also these arguments borderline pathetic... something is detected not to your likes? Then just add it to the exclusion list. No, instead... you junk the entire software.

Its faster to say... I just hate anything other than AVAST... more credible.

 

Yes one month back. But you can't compare the 1 to 120, this is nonsense. Because you compare total different things: different tests, websites vs. files, different FP-test-set sizes...

 

Is it? Since when is, i don't know, MS Office package somehow different between FP test in real-world testa nd file detection test? Did it somehow repack itself with weird packers and got stripped of the digital signatures? Got files somehow modified taht they magically dropped from the avast! whitelists? Or any other software used in the false positives tests? How can it only detect 1 false positive in the entire year 2014 of Real-World tests yet it detects 120+ in each of last two file detection tests? And what's even more laughable, people diss it just because they saw this sensational title "avast! had ridiculous amount of FP's", but they just don't give a F about absolutely spotless real-world tests track record through the ENTIRE year 2014. Weren't like nearly all of you saying other tests don't matter anymore, real-world is the one that actually matters? But now, it somehow not important. Weird logic. I'm not defending avast! because it's my favorite, i'm defending it based on facts. You can't be among the very best in false positives and in other test that goes in parallel the worst. It just makes zero sense and just calls for seriously flawed testing methodology. Period. If you make some sort of informative tests, they have to be the kind that reflects the actual state and not great in one and the worst possible in other for the same bloody part of the year. How does such tests help users to decide when it basically contradicts itself? If the tests were actually meaningful, the false positive scores would generally be roughly within margin of error within the same month. Yet they are galaxies apart. Just doesn't make any sense. Either drop real-world false positive test or the file scanning one, because apparently one of these two is pointless and gives iditoic results.

 

READ the methodology on AV-C website, then you will see that this are different things.
- file detection test: FP-testset of clean files. Helps to to detect "signature problems" f.e. with evo-gen
- real-world-detection test: live-links...

Beside that: you can't compare 1 to 120, when the basis is completely different. Read something about "base rate fallacy" for that.
__

It's just one test, so why not take it as just one indicator for a problem with signature scanning ?? Why attacking the test?

 

Because it, erm, makes no sense?

- file detection test: FP-testset of clean files. Helps to to detect "signature problems" f.e. with evo-gen
- real-world-detection test: live-links...

So, first one checks the signatures part and the real-world test uses magic pixies to perform file scaning or what? If anything, real-world test should have more false positives since more detection capabilities are deployed there. Yet it's the other way around...

 

F-Secure will be happy to discuss the same false-positives case: 9 vs 66 instances for the same month in both tests

 

Avast has never been a top tier free AV. Always seems to be in the upper middle category rather than top one.

 

FWIW, here's a less mentioned ranking service to throw on the pile with the rest: https://threatcenter.crdf.fr/?Stats

Their results, although generated daily, combine A-V with A-M products and therefor are problematic.

Enjoy

 

I really think you should READ the methodology and after that something about the "base rate fallacy" (Kahnemann wrote some good papers here) Otherwise further discussion makes no sense, cause it's based on completely wrong assumptions from your side.

And I know from you, that you in general have a good and deep unterstanding and are a critical mind! But in this topic your posts somehow seem to be nothing more than baseless bashing.

 

Sure, whatever you say...

 

I second that.

 

Third it.

Not sure why people are up in arms with below average or average results from Avast. Isn't that normal? I'm just excited Trend 2015 is scoring 100% on all of the tests now with the Deep Security Infrastructure behind it. A lot of AV companies talk up new technology, and use fancy new names for it, but this one is real.

 

If I recall, 2-3 years back, Avast was an awesome, top tier free AV. Then they started added a lot of new features that sounded awesome on paper, yet protection seemed to drop off.

 

Avast always had lot of features as compared to other free offers but still detection is never extra ordinary.

In spite of that I don't find any other decent option except Avast to install for my friends who want a free AV.

 

I used to recommend Avast for client computers 3-4 years ago. But with more and more junk added to Avast along with more annoyances, plus the gradual decline of detection rates over the past few years. Therefore, I have had to recommend other solutions. I understand that this recent test of Avast showing a crazy number of false positives is likely just a one-off situation that I'm certain the Avast devs will resolve quickly.

 

Agree with that.

So what are you using now for this purpose.

 

Same here.

Now I recommend Q360IS + Comodo firewall, cruelsister's config. No more avast for me.

If my friends can read chinese, I recommend them the chinese version of 360.

 

Being in the field, one thing that concerns me is that in communist, or totalitarian countries they require firms to implement backdoors, and if necessary detailed information gathering. Kaspersky for example is only allowed to exist because he works with Russian Security on-demand, and he's said this himself. I believe in China it's the same. By contrast in North America much of our protections come from corporations worried about exposure, and the financial damage it can cause, also the legality of it so opt-out is often provided. So I avoid RU and CN products in terms of security solutions for this reason. I found programmed exploits in a couple of Chinese browsers, and when I questioned their developers I was told these are 'required' by the Chinese Security Services.

Something to consider at least.

 

Yes I agree with you about backdoors being planted in some software. But if I were an American citizen I would prefer Chinese government getting info about me than my own government getting that same info. It's not like Chinese police will come knocking on my door.

 

Agreed, but it's very likely your information could be harvested for Chinese corporations for datamining, corporate secrets, etc. We are prohibited where I work from deploying RU/CN solutions to corporate clients for this very reason. While the feds can come knocking, they are much less likely to sell your information to state sponsored data miners, or to take your business secrets. It would be corporate suicide for a company like Symantec to be implicated in such schemes.