PrevX test by MRG

[ludo021288]

Here is a very recent test published today by MRG. It compared PrevX to other cloud antiviruses like Bluepoint, Hitman Pro, Panda Cloud and Immunet Protect.
Seems like PrevX didn't score very well this time :(

http://malwareresearchgroup.com/forum/viewtopic.php?f=20&t=165

[firzen771]

ive never put much faith in MRG so i wuldnt take this test to seriously IMO.

EDIT: if u read on it says they only ran on-demand tests and did not execute the malware, well now that completely explains why prevx scored so low.

[the Tester]

That test was posted in this thread (510) by erikloman from Hitman Pro.


http://www.wilderssecurity.com/showthread.php?t=236732&page=21

[funkydude]

Not many people care for MRG test results.

[Triple Helix]

Was it these guys that got caught with many user names on the Forums and were outed to be dishonest by the Forum Admins? :wacko:

TH

[funkydude]

-{ Quote: "Was it these guys that got caught with many user names on the Forums and were outed to be dishonest by the Forum Admins? :wacko:
" }-

Yes. ::)

[Triple Helix]

Thought so! Well the test mean nothing then! :thumbd:

TH

[papillonn]

Let me know are there any reliable cloud tests?

[Triple Helix]

-{ Quote: "Let me know are there any reliable cloud tests?" }-

I guess we will be still waiting then!

TH

[PrevxHelp]

We haven't gotten any details regarding the test samples so we honestly have no idea if they actually are malicious.

However, as firzen771 pointed out, Prevx's right-click scanner provides only a very small fraction of the detection/protection which Prevx fully provides. Testers trying to test Prevx on-demand are incorrect in their methodology and clearly don't understand how Prevx works :-\

[papillonn]

-{ Quote: "clearly don't understand how Prevx works" }-

It is a shame for a testing group obviously.

[Defenestration]

MRG have responded:

http://malwareresearchgroup.com/forum/viewtopic.php?f=20&t=165#p515

[Scoobs72]

-{ Quote: "MRG have responded:

http://malwareresearchgroup.com/forum/viewtopic.php?f=20&t=165#p515" }-

The response is typical of this bunch of people. For a group that has an extremely tarnished reputation (and I'm being generous with that) they behave as if they are kings of the hill.

[Triple Helix]

As I said dishonest and now untrustworthy these guys are full of BS >:( Foot in mouth syndrome!

TH

[Triple Helix]

-{ Quote: "MRG have responded:

http://malwareresearchgroup.com/forum/viewtopic.php?f=20&t=165#p515" }-

Sorry forgot to say thanks for the update!

TH

[trjam]

and just what if, they are right. Prevx had intended from day one to get reviewed by PCMag. It took a few months but in the end they got what they wanted. Right or wrong? I dont know but is it any differnet then Symantec and their paid results. I really dont trust any testing vendors anymore, period. The best tests are the ones that members here report based on fact, not dollars.

[Defenestration]

-{ Quote: "MRG have responded:

http://malwareresearchgroup.com/forum/viewtopic.php?f=20&t=165#p515" }-As they mention at the beginning though, they only tested on-demand, whereas Prevx's strength lies in it's ability to detect executed/running malware in real-time, which I guess is why Prevx scores so low.

[the Tester]

-{ Quote: "and just what if, they are right. Prevx had intended from day one to get reviewed by PCMag. It took a few months but in the end they got what they wanted. Right or wrong? I dont know but is it any differnet then Symantec and their paid results. I really dont trust any testing vendors anymore, period. The best tests are the ones that members here report based on fact, not dollars." }-


I agree 100%.
Paid results are nothing more than advertisements regardless of who the vendor is.

[demoneye]

hi

if MRG, the testing company "doesnt know how prevx operate, and there for tests are incorrect"("official prevx reply Testers trying to test Prevx on-demand are incorrect in their methodology and clearly don't understand how Prevx works") if so , how can a common user know how to use prevx full power or even in the correct way?;D

[JRViejo]

Removed two Off-Topic posts. Lets keep the discussion focused on PrevX and please stay on topic. Thanks!

JR

[Vikorr]

-{ Quote: "and just what if, they are right. Prevx had intended from day one to get reviewed by PCMag. It took a few months but in the end they got what they wanted. Right or wrong? I dont know but is it any differnet then Symantec and their paid results. I really dont trust any testing vendors anymore, period. The best tests are the ones that members here report based on fact, not dollars." }-

Doesn't Prevx do behavioural analysis? If there is very new malware installed and all you do is a system scan (whether full or part), and not have the malware run (ie give Prevx no behaviour to analyse), then it's no surprise that Prevx only picks up 29%.

MRG say they don't give the products a chance to do behavioural analysis, and this to me seems to be what Prevx says when they say 'you guys don't understand how Prevx works'. MRG seems to have missed this point entirely.

Of course, there are also settings on Prevx that allow for the detection of new files, not seen by Prevx, and for the test, I would doubt that MRG had them enabled (as of course, by default they aren't set so sensitive).

Then there is the speed at which detections are added.

Yeah, it's hard to judge from their test, but from my admittedly limited knowledge of Prevx (which seems to be more than MRG) it does seem they employed a flawed test.

[ako]

The test seems OK. It was an on-demand test. Period. (How do you perform an on-demand scan with Hitman pro, by the way? It has been spesifically built to find living malware.)

The statistics is of course rather small (and maybe somehow biased), but the two first tests are rather consistent (MSE being only anomaly):

1) Bluepoint Security = 61.6%
2) Hitman Pro = 57.0%
3) A-Squared = 52.2%
4) AntiVir = 51.6%
5) Microsoft Security Essentials 45.4%
6) NOD32 = 33.8%
7) Prevx = 21.0%
8) Panda Cloud = 13.0%*
9) Immunet Protect = 5.2%*

1) Hitman Pro = 66%
2) Bluepoint Security = 60%
3) Prevx = 29%
4) Panda Cloud = 15%*
5) Immunet Protect = 7%*

Microsoft Security Essentials 24%.

[ako]

A year ago from a similar test or "test" (http://winnow.oitc.com/avmalwarestats.php):

AntiVir = 62%
NOD32 = 34%
Prevx = 33%
Microsoft 25%

These results are not too far from the one above.

This is not to say I think the test is relevant for measuring the real-time protection Prevx gives, see eg. my own small test:
http://www.wilderssecurity.com/showthread.php?t=252928

[Triple Helix]

ako Please read this thread http://www.wilderssecurity.com/showthread.php?t=251113

Thanks,

TH

[PrevxHelp]

In the interest of not fueling further discussions on topics about a group which have been historically closed by other moderators on the forum, I'm going to continue and close this thread as well.

We recommend continuing discussions over on the Malware Research Group's own forums.