Av-Comparatives: On-Demand Comparative August 2010

Discussion in 'other anti-virus software' started by hckyo, Sep 23, 2010.

Thread Status:
Not open for further replies.
  1. drakester

    drakester Registered Member

    Joined:
    Feb 17, 2010
    Posts:
    54
    A real way to prove a point. Getting real results out there. :cool:

    avast! is very good. :thumb:
    Dissapointed on Kaspersky, not only their latest version has issues but the FPs have increased massively.
    Norton did well in my opinion.

    All these tests show that no one is flawless, gotta keep a layered approach and don't put all your bets in a single vendor.
     
  2. guest

    guest Guest

    False Positive for Avast: 9
    False Positive for Bitdefender: 4
    Avast+Bitdefender FP=13

    FP for Gdata=Avast+Bitdefender+Internal Whitelist Database= 15

    How is it possible?
     
  3. vlk

    vlk AV Expert

    Joined:
    Dec 26, 2002
    Posts:
    621
    That's because Gdata uses different versions of the engines than the ones used by the standalone products (e.g. it uses avast 4.x engine while the test was done with avast 5... not sure about BitDefender).

    Thanks
    Vlk
     
  4. ReverseGear

    ReverseGear Guest

    As mentioned earlier MSE juts needs to improve its detection just a little bit and it will give all other freewares a run for there money
    avast did better den my expectation...
    if avira free would have been used what would be d difference any guesses ?
     
  5. AvinashR

    AvinashR Registered Member

    Joined:
    Dec 26, 2009
    Posts:
    2,063
    Location:
    New Delhi Metallo β-Lactamase 1
    An awesome result for Panda Labs and Avast ..
     
  6. guest

    guest Guest

    Comodo have never refused to be tested.
    Seems that is not easy get into, maybe AVC has many clients now.
     
  7. AnonOT

    AnonOT Registered Member

    Joined:
    Feb 1, 2010
    Posts:
    16
    finally great results for Avast
    Avira staying strong :)
     
  8. IBK

    IBK AV Expert

    Joined:
    Dec 22, 2003
    Posts:
    1,886
    Location:
    Innsbruck (Austria)
    Please avoid asking if we tested products x etc. Due NDA's, we can not disclose which products are tested internally etc. If they get published is up to the vendors.
     
  9. progress

    progress Guest

    98 FPs :eek: It's dangerous to use Panda ... :doubt:
     
  10. Matthijs5nl

    Matthijs5nl Guest

    To be honest it is deadly, since the rest of Panda is great, (detection and performance). I also thought I heard Panda a lot about how they improved their fp rate, you can't find a trace of it in this test though.
    Panda has more problems though (in the cloud version the slow icon loading, shitty restore/quarantine/remove behavior, and in all versions an extremely slowly loading GUI and messages etc.). I wish they would concentrate on that for a while. Would be a good step to really become a tier-1 product.

    Dissapointed by Kaspersky again, happy with my ESET's result. Also Avira did great and according to this test I have to reflect my previously taken standpoint regaring avast.
     
  11. Leo2005

    Leo2005 Registered Member

    Joined:
    May 31, 2007
    Posts:
    179
    Location:
    Braunschweig (Germany)
    There wouldn't be a diffrence. This is an on demand test. as the free version uses the same signatures and engine it would be the same result.
     
  12. PJC

    PJC Very Frequent Poster

    Joined:
    Feb 17, 2010
    Posts:
    2,959
    Location:
    Internet
    AV-Comparatives: On-Demand Comparative (August 2010)

    -Avira Remains a Top performer!
    -avast! becomes better and better!
    -Kaspersky tied with AVG :D (= in Detection, because in FPs...:thumbd:)
     
  13. ReverseGear

    ReverseGear Guest

    wouldnt the proactive module in premium version come into play ?
     
  14. King Grub

    King Grub Registered Member

    Joined:
    Sep 12, 2006
    Posts:
    818
    The proactive module is supposed to kick in upon execution, not on-demand scanning.
     
  15. progress

    progress Guest

    However - On-Demand tests are totally outdated, it would be better to do more protection tests against unknown malware :thumb:
     
  16. ReverseGear

    ReverseGear Guest

    ohh ok..thnx for explaining
     
  17. CogitoTesting

    CogitoTesting Registered Member

    Joined:
    Jul 4, 2009
    Posts:
    901
    Location:
    Sea of Tranquility, Luna
    Your statement simply does not match the facts here. The only thing you have to do is to go to their forum and review all the threads relevant to CIS and testing and behold for yourself. Melih has consistently been adamant in all of his excuses for CIS NOT to be tested. Really it has been one pathetic excuse after another.

    Thanks.
     
    Last edited: Sep 24, 2010
  18. RejZoR

    RejZoR Lurker

    Joined:
    May 31, 2004
    Posts:
    6,426
    No they are not outdated, you just have to understand them (this is the point where like 90% of ppl fail to understand what they mean). This test will be outdated when ALL programs without single exception will detect 99,9%. Until then, the test is very valid.
     
  19. progress

    progress Guest

    Do you really think that there are only 900.000 samples out there? :rolleyes: I can't believe it ...

    I could create 20 samples today and your 99,9% scanner won't detect them :)
     
  20. Matthijs5nl

    Matthijs5nl Guest

    I also still think this kind of tests are still relevant, I totally agree with Rejzor.

    There is always a lot of chitchat about zero-day malware etc; but people forget that only a very small amount will actually target people (with small amount a mean a rate in the single digits). Around 80% won't survive the first 24 hours, a lot are actually the same malwaresamples while people (like languy) think they got "different" threats (like hosted on a different site/slightly repacked etc). Actually all the "zero-day" malware people talk about are in reality >24 hours (otherwise they wouldn't have survived that long to make it on MDL or whatever kind of forum). Also real zero-day malware ain't the biggest threat to people (just the latest variant of real malware like the latest Zbot etc.), since the zero-day malware ain't present on the good side of the web.
    If one of those zero-day malware samples does infect people anyway, it will only infect <10 pc's in the total world. Conclusion: zero-day malware is overrated.

    Relevant for people who don't go searching for malware, so normal people: an up-to-date-pc (7/Vista), some pc/internet knowledge, and an up-to-date antivirus (one of the big ones, so there are like 10 alternatives (all probably equally good for normal people)) will serve any person well. And exactly for this reason this "outdated" test is still relevant.
    That is also why HIPS and virtualization and so on don't really get more popular, for normal people who got medium pc knowledge malware ain't a really big deal.
    Behavior and reputation-based is a good approach in my eyes, together with the "old" generic signatures and heuristics (basically also Norton's approach).
     
    Last edited by a moderator: Sep 24, 2010
  21. progress

    progress Guest

    <10 pc's in the total world? :doubt:

    Edit: The Effectiveness of Antivirus on New Malware Samples
     
    Last edited by a moderator: Sep 24, 2010
  22. Osaban

    Osaban Registered Member

    Joined:
    Apr 11, 2005
    Posts:
    5,614
    Location:
    Milan and Seoul
    I can't believe it myself when I think about it, but in 5 years I've been monitoring my computers, browsing has created a few alerts, almost negligible within this span of time.

    On the other hand plugging USB flash drives from strangers into my machines (a common occurrence in my job) has been a potential calamity: I'm not kidding, but first with Eset and then Avira, I think they have detected more than 100 different signatures. It really depends on one's exposure vectors, but malware is out there, it is only a matter of how it will make its way into your computer.
     
  23. Matthijs5nl

    Matthijs5nl Guest

    Had to make an edit: I meant that one of THOSE so called zero-day malware samples will only infect like 10 pc's.

    They collected 152 fairly new samples over a few days. I wonder how many pc's really have been infected by that malware samples at the time that test was done, I am quite sure it will be "surprisingly" low.
     
  24. guest

    guest Guest

    I guess that you have not read the forums for a long time, you can find even 2 years old AV tests of Comodo
    And the most recent in July 2010: http://www.pcsecuritylabs.net/document/report/PCSL_Total_Protection_Test_July_2010_EN.pdf
    Feb 2010: http://virusinfo.info/index.php?page=testseng
    http://www.anti-malware-test.com/?q=node/93/
    Sep 2009: http://www.virus.gr/portal/en/content/2009-08,-10-august-05-september

    Comodo have not been tested yet by AVC as many other products but they have been trying to get into, maybe is one of the products tested "in secret".
    I don't know which are the requirements or limitations that AVC has in order to test more products at the same time.

    If you remove AVC there aren't many test that are done periodically with public reports.

    Instead to said something without any fundament tell me please when Comodo said that they dont want to be tested.

    Simply I dont belive that Comodo does not want their product to be tested, is my opinion.
     
    Last edited by a moderator: Sep 24, 2010
  25. clocks

    clocks Registered Member

    Joined:
    Aug 25, 2007
    Posts:
    2,787
    I suspect Comodo was tested, didn't like the results, and asked AVC not to report them. Melih promised many months ago that CIS would be in this round. There was plenty of time for them to be tested. Yesterday, after the results were released, Melih started his "detection rates don't matter" campaign. Something else is going on here.
     
Thread Status:
Not open for further replies.
  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.