Malware-Test Lab: Antivirus Comparison Report (February 26, 2007)

Discussion in 'other anti-virus software' started by sai7sai, Feb 26, 2007.

Thread Status:
Not open for further replies.
  1. sai7sai

    sai7sai Registered Member

    Joined:
    May 3, 2006
    Posts:
    21
    Location:
    Taiwan
    Detection rate for different antivirus softwares (20070226):

    Total Malware = 190,638

    ‧Fortinet FortiClient Host Security 3.0: 94.70%
    ‧BitDefender Internet Security v10: 92.36%
    ‧AOL Active Virus Shield (Free): 89.02%
    ‧AVIRA AntiVir PersonalEdition Classic 7.0 (Free): 88.44%
    ‧GRISOFT AVG Internet Security 7.5: 88.37%
    ‧BitDefender 8 Free Edition (Free): 86.52%
    ‧Rising Antivirus 2007: 84.78%
    ‧Kaspersky Internet Security 6.0: 84.77%
    ‧Panda Internet Security 2007: 79.16%
    ‧MicroWorld eScan Internet Security 8.0: 77.68%

    ‧ClamWin Free Antivirus 0.90 (Free): 76.57%
    ‧Sunbelt CounterSpy V2: 76.39%
    ‧ALWIL avast Professional 4.7: 73.41%
    ‧F-Prot Antivirus 3.16f: 72.75%
    ‧Filseclab Twister Anti-TrojanVirus V7: 62.66%
    ‧Trend Micro Internet Security 2007: 60.53%
    ‧CA Internet Security 2007: 58.61%
    ‧AhnLab V3 Internet Security 2007 Platinum: 56.44%
    ‧Kingsoft Internet Security 2007: 55.55%
    ‧ESET NOD32 2.7: 49.17%
    ‧Jiangmin Antivirus KV 2007: 45.16%
    ‧Webroot Spy Sweeper with Antivirus: 40.71%
    ‧CAT Quick Heal Total Security 2006: 37.29%
    ‧Hauri Virobot Desktop 5.0: 34.38%
    ‧Comodo Antivirus 1.1 (Free): 9.92%

    For the detailed information, please refer to HERE and read attached file.

    Note: due to update schedule of virus signature is different, so the admissible error is between -4.3% and +4.3%.
     
  2. C.S.J

    C.S.J Massive Poster

    Joined:
    Oct 16, 2006
    Posts:
    5,029
    Location:
    this forum is biased!
    doesnt add up,

    i honestly believe they are working out their percentages wrong, by using files scanned minus infected files, which just wont add up.
     
  3. lucas1985

    lucas1985 Retired Moderator

    Joined:
    Nov 9, 2006
    Posts:
    4,047
    Location:
    France, May 1968
    This must be a bad joke :rolleyes:
     
  4. sai7sai

    sai7sai Registered Member

    Joined:
    May 3, 2006
    Posts:
    21
    Location:
    Taiwan
    Why do you think it is wrong? How do you calculate detection rate?
     
  5. hbkh

    hbkh Registered Member

    Joined:
    Jan 15, 2004
    Posts:
    128
    Location:
    Ohio, USA
    In what world could Rising or AVG beat Kaspersky? Bad test. :(
     
  6. ASpace

    ASpace Guest

    Yes , how the hell AOL AVS has detection rate 89% but Kaspersky itself has 84% ?? 5% is not little when we have ~190000 threats tests, it is even not little when the same signatures are used

    How can Rising AV get better result than NOD32,F-prot,AVG,Panda,(even KAV by 0.01%) ?

    What is Filseclab o_O

    This is ridiculous :thumb:
     
  7. sai7sai

    sai7sai Registered Member

    Joined:
    May 3, 2006
    Posts:
    21
    Location:
    Taiwan
    You forgot one thing "due to update schedule of virus signature is different, so the admissible error is between -4.3% and +4.3%.". You can see AOL Active Virus Shield got a better rank, but it is almost the same as Kaspersky.
     
  8. Stefan Kurtzhals

    Stefan Kurtzhals AV Expert

    Joined:
    Sep 30, 2003
    Posts:
    701
    I don't wanna know how much trash this collection contains and if the settings for various antivirus products were set to non-optimal detection by intent or just because the tester is clueless. :rolleyes:
     
  9. plantextract

    plantextract Registered Member

    Joined:
    Feb 13, 2007
    Posts:
    392
    and the marmote wrapped the chocolate in aluminum foil.
    You got to be joking kav and avs have the same engine & same signatures,

    also, nod32 only 48% what the heck is that.

    so when were both tests taken, there's a 8200 file difference between those.
    That's not how you do tests. if you can't run the simultaneously (on different pcs of course), install them all, update, make images of that and then use those images on a scan so everybody gets a fair advantage.
     
  10. IBK

    IBK AV Expert

    Joined:
    Dec 22, 2003
    Posts:
    1,819
    Location:
    Innsbruck (Austria)
    or because DEMO/evaluation versions were used
     
  11. sai7sai

    sai7sai Registered Member

    Joined:
    May 3, 2006
    Posts:
    21
    Location:
    Taiwan
    Filseclab is a china antivirus company (http://www.filseclab.com/eng/default.htm).

    For detection rate, only for reference, you can choose any antivirus software you like. The most important things are scan performance and product quality.
     
  12. IBK

    IBK AV Expert

    Joined:
    Dec 22, 2003
    Posts:
    1,819
    Location:
    Innsbruck (Austria)
    escan and kav should score more or less same, so the difference is at least 7,1%
     
  13. sai7sai

    sai7sai Registered Member

    Joined:
    May 3, 2006
    Posts:
    21
    Location:
    Taiwan
    We are antivirus/virus/testing experts, don't be doubtful. If antivirus softwares (trial/evaluated version) don't include full functionalities, we wouldn't test them.
     
  14. Smokey

    Smokey Registered Member

    Joined:
    Apr 1, 2002
    Posts:
    1,513
    Location:
    Annie's Pub
    Let's keep it short and painless: the ehhh 'test' itself is trash.
    I can't imagine that even a person with less then common sense will put his faith and trust in this so called ehhhh 'test'.
    To me it look like a bad joke:rolleyes:

    Regards,

    Smokey
     
  15. plantextract

    plantextract Registered Member

    Joined:
    Feb 13, 2007
    Posts:
    392
    from the website:
    6. The following antivirus softwares’ scan performance is very slow.
    ...
    ‧AVIRA AntiVir ProfessionalEdition Classic 7.0 (Free) (over 64 hours)
    ...


    wow, if AVIRA is slow, then a lot of the others would be standing still, maybe even go to negative percentages in terms of completion.

    Experts, you don't say, kind of hard to belive.
     
  16. sai7sai

    sai7sai Registered Member

    Joined:
    May 3, 2006
    Posts:
    21
    Location:
    Taiwan
    AOL AVS should be the same as Kaspersky. In the future, we can test them at the same time.

    For NOD32, we are very surprising too.
     
  17. Londonbeat

    Londonbeat Registered Member

    Joined:
    Sep 21, 2006
    Posts:
    350
    Do you know why AOL got 5% higher than KIS 6, when they use the same engine? :doubt: Were all products signatures updated at the same time?

    Londonbeat
     
  18. Stefan Kurtzhals

    Stefan Kurtzhals AV Expert

    Joined:
    Sep 30, 2003
    Posts:
    701
    You can say quite some bad things about Avira, but I never heard someone saying that the scan speed is slow. Funny, indeed. :rolleyes:

    They obviously didn't turn on the Adv. Heuristic for NOD32 and I bet they didn't turn on the Avira heuristic or left it on "normal" level.
     
  19. ASpace

    ASpace Guest

    Then why do create tests with old versions . F-prot 3.16f is quite old version if tests are from 20070226
     
  20. Rico

    Rico Registered Member

    Joined:
    Aug 19, 2004
    Posts:
    1,701
    Location:
    Texas
    Hmmmmmmm,

    The folks doing the testing are from 'Trend Micro' so it looks like there not biased.

    Take Care
    Rico
     
  21. sai7sai

    sai7sai Registered Member

    Joined:
    May 3, 2006
    Posts:
    21
    Location:
    Taiwan
    antivir_personaledition_scan_time.png
     
  22. plantextract

    plantextract Registered Member

    Joined:
    Feb 13, 2007
    Posts:
    392
  23. plantextract

    plantextract Registered Member

    Joined:
    Feb 13, 2007
    Posts:
    392
    wow, thanks.
    btw.
    173202 detections 190638 files, so
    173202 ----- 190638
    x ---------- 100
    x=173202*100/190638~=90,85


    that's math :D
     
  24. IBK

    IBK AV Expert

    Joined:
    Dec 22, 2003
    Posts:
    1,819
    Location:
    Innsbruck (Austria)
    you can not rely on the detection numbers that AV programs give in their results.
     
  25. Londonbeat

    Londonbeat Registered Member

    Joined:
    Sep 21, 2006
    Posts:
    350
    Honeypot malware used - how did you make sure samples were not corrupted?

    Some AV's (NOD32) were tested with signatures from 16/2/07 whereas others are tested with signatures from 23/2/07?
     
Loading...
Thread Status:
Not open for further replies.