AV-Comparatives False Alarm Test March 2014

Discussion in 'other anti-virus software' started by MrBrian, Apr 24, 2014.

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

    MrBrian Registered Member

    Joined:
    Feb 24, 2008
    Posts:
    6,032
    Location:
    USA
    http://www.av-comparatives.org/false-alarm-tests/

    Background info from Wikipedia:
     
    Last edited: Apr 24, 2014
  2. anon

    anon Registered Member

    Joined:
    Dec 27, 2012
    Posts:
    4,094
  3. MrBrian

    MrBrian Registered Member

    Joined:
    Feb 24, 2008
    Posts:
    6,032
    Location:
    USA
    Noted. This is a different test than the title of that thread though. IMHO, when there are separate but related items, the best thing to do is have a separate thread for each item, and if desired post hyperlinks to those related threads.
     
    Last edited: Apr 24, 2014
  4. xxJackxx

    xxJackxx Registered Member

    Joined:
    Oct 23, 2008
    Posts:
    4,047
    Location:
    USA
    It doesn't look like anyone is too interested in false positives... They have caused me more problems over the years than actual malware and viruses. I find the results to be pretty consistent with my experiences.
     
  5. Minimalist

    Minimalist Registered Member

    Joined:
    Jan 6, 2014
    Posts:
    5,062
    +1. In some cases FPs can be more dangerous than viruses.
     
  6. Impet

    Impet Registered Member

    Joined:
    May 5, 2013
    Posts:
    894
    Avast 94 false positives? Well done ... :thumbd: But hey, you get grime fighter, toolbars, chrome and other junk for free. :eek:
     
  7. RejZoR

    RejZoR Registered Member

    Joined:
    May 31, 2004
    Posts:
    6,426
    You can mock it all you want because of this test, but out of all AV's, avast! is among rare few that has by far the least false positives in general. And my systems eat a lot of data ranging from very old stuff to super brand new apps. So far i think i only had 1 Evo-Gen false positive since it's introduction. And that's a very low number if you think of it...
     
  8. aztony

    aztony Registered Member

    Joined:
    Sep 9, 2012
    Posts:
    547
    Location:
    USA Southwest
    Baidu had a horrendous FP rate when I trial-ed it back when it was 1st released, what, a year ago. Seems they haven't improved much since then either.
     
  9. anon

    anon Registered Member

    Joined:
    Dec 27, 2012
    Posts:
    4,094
    Avast FPs are ridiculously high
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2014&month=3&sort=1
    but:

    False-Alarm-Test March 2014 PDF, page 9:
    AV-C Remark=
    "Avast had 94 false alarms. Win32:Evo-gen is a relatively new heuristic detection of Avast, it is
    expected that the number of false alarms caused by it will be much lower in future."
    http://www.av-comparatives.org/wp-content/uploads/2014/04/avc_fps_201403_en.pdf
     
  10. Impet

    Impet Registered Member

    Joined:
    May 5, 2013
    Posts:
    894
    Thank you, I see but detection is still mediocre. Nearly all other competitors had better detection. Let's see if Avast is able to fine-tune Evo-gen for the next test. :)
     
  11. anon

    anon Registered Member

    Joined:
    Dec 27, 2012
    Posts:
    4,094
    Your first comment was about FPs. -> You got an answer re the FPs.
    In reply, instead of the FPs now you are talking about the low detection rates!
    lol!
     
  12. Impet

    Impet Registered Member

    Joined:
    May 5, 2013
    Posts:
    894
    I thought more FPs means better detection as well, as we can see my conclusion was false. o_O
     
  13. Thankful

    Thankful Savings Monitor

    Joined:
    Feb 28, 2005
    Posts:
    3,736
    Location:
    New York City
    Emsisoft FPs courtesy of Bitdefender.
     
  14. RejZoR

    RejZoR Registered Member

    Joined:
    May 31, 2004
    Posts:
    6,426
    Like i said before, you're greatly exaggerating it. Or shall i say, the test is exaggerating it. With these scores, one would think you'd get false positives on like every 3rd file in existence. But reality is, i got far more false positives with Qihoo. Or Comodo. I never even had much problems with hardened Mode (Aggressive) and that one basically prevents everything that isn't whitelisted.
     
  15. anon

    anon Registered Member

    Joined:
    Dec 27, 2012
    Posts:
    4,094
    Here we have the AV-C test results and we make comments on that specific results.
    "RejZoR's reality" / "anon's reality", etc it's another story, IMHO.
     
  16. RejZoR

    RejZoR Registered Member

    Joined:
    May 31, 2004
    Posts:
    6,426
    You also don't see massive floods of false positive screams on avast! forums or here or anywhere else. Do the math on those "realities"...
     
  17. clocks

    clocks Registered Member

    Joined:
    Aug 25, 2007
    Posts:
    2,559
    After not using it for a long time, I recently tried Norton IS. While I mostly liked it, and found it very light, it gave me too many issues with FPs.
     
  18. hawki

    hawki Registered Member

    Joined:
    Dec 17, 2008
    Posts:
    1,955
    Location:
    DC Metro Area
  19. Noob

    Noob Registered Member

    Joined:
    Nov 6, 2009
    Posts:
    6,468
    Wow that Evo-Gen really messed up Avast! results . . . Didnt Avast! release Evo-Gen a couple months ago? If so, it shouldnt have so many FP's because there has been enought time.
     
    Last edited: Apr 27, 2014
  20. anon

    anon Registered Member

    Joined:
    Dec 27, 2012
    Posts:
    4,094
    FP Test, Sept 2013 = Avast: 10 FPs
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2013&month=9&sort=1

    FP Test, March 2013 = Avast: 14 FPs
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2013&month=3&sort=1

    FP Test, Oct 2012 = Avast: 11 FPs
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2012&month=10&sort=1

    FP Test, March 2012 = Avast: 14 FPs
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2012&month=3&sort=1

    FP Test, August 2011 = Avast: 10 FPs
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2011&month=8&sort=1

    FP Test, Feb 2011 = Avast: 19 FPs
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2011&month=2&sort=1

    FP Test, August 2010 = Avast: 9 FPs
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2010&month=8&sort=1

    FP Test, Feb 2010 = Avast: 11 FPs
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2010&month=2&sort=1

    FP Test, Feb 2009 = Avast: 28 FPs
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2009&month=2&sort=1

    FP Test, August 2009 = Avast: 5 FPs
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2009&month=8&sort=1

    FP Test, August 2008 = Avast: 47 FPs
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2008&month=8&sort=1
    -------------------------
    FP Test, March 2014 = Avast: 95 FPs
    http://chart.av-comparatives.org/chart1.php?chart=chart6&year=2014&month=3&sort=1

    False-Alarm-Test March 2014 PDF, page 9:
    AV-C Remark=
    "Avast had 94 false alarms. Win32:Evo-gen is a relatively new heuristic detection of Avast, it is expected that the number of false alarms caused by it will be much lower in future."
    http://www.av-comparatives.org/wp-content/uploads/2014/04/avc_fps_201403_en.pdf
     
    Last edited: Apr 27, 2014
  21. Feandur

    Feandur Registered Member

    Joined:
    Jun 15, 2005
    Posts:
    401
    Location:
    Australia
    @anon: RE: You also don't see massive floods of false positive....

    Over the 2008 - 2014 data @ your post above-

    Kaspersky had a "bad hair day' in 2010 with 46 fp's and in 2008 with 28 pf's;
    Avira had a "bad hair day" in 2009 with 24 fp's, followed by 21 in 2009 and 20 in 2013;
    Avast had a "bad hair day" in 2008 (47), 2009 (28 ) and recently in 2014 (95).

    However, more representative of the ability (or usual result, rather than the exceptional result) of the AV is an average range, ignoring the 2 lowest and 2 highest outliers in the stats....

    Kaspersky is more typically 5 - 14 fp's;
    Avira is more typically 9 - 20 fp's;
    Avast is more typically 10 - 28 fp's

    So, I rank the AV's in that descending order of merit, (Kasp', Avira, Avast), from a pf's point-of-view.

    Have been a user of Avira in the past and liked it.

    Now with EAM, and make no apologies for that.
     
  22. anon

    anon Registered Member

    Joined:
    Dec 27, 2012
    Posts:
    4,094
    Is that a specific rule on statistics? Why not the 1 or the 3, 4, 5, x H/L outliers in the stats?

    For me, what it counts is the present results and the AV capabilities. Not the past or the average values.
     
  23. Minimalist

    Minimalist Registered Member

    Joined:
    Jan 6, 2014
    Posts:
    5,062
    ESET has a good track history. Trend is good also. :thumb:

    upload_2014-4-27_9-8-14.png
     
  24. Minimalist

    Minimalist Registered Member

    Joined:
    Jan 6, 2014
    Posts:
    5,062
    Made comparison with some other AVs. Good news - most of them have declining trend :thumb:

    upload_2014-4-27_9-39-12.png
     
  25. Feandur

    Feandur Registered Member

    Joined:
    Jun 15, 2005
    Posts:
    401
    Location:
    Australia
    Ok anon. I'll bite...

    Provocative as always...and your hubris breathtaking.

    Bell shaped curve, remove the extreme outliers, to get a general sense of the mean and standard deviation. Gut feel only..standard practice in various engineering arts, eg: road pavement technology when interpreting FWD pavement responses.

    Go back and do your engineering degree again.... and I'm speaking engineering practice, not book learning.

    < Yes, I know you'll spend the next 30 seconds looking up the pedantic meaning of standard deviation to 5 decimal places, to shove it up my a**. Could'nt give a stuff! >

    The present instantly becomes the past . You know that. Product development sometimes happens with hickups..."bad hair days"....Stop quoting results 2013, 2012, 2011, 2010, 2009, 2008.

    Like you, I'll ignore the good reason why someone might mention past data.

    shove off.

    F
     
Loading...
Thread Status:
Not open for further replies.