AV-Comparatives - Real-World Protection Test - April 2013

New chart up:

http://chart.av-comparatives.org/chart1.php

One product reached 100% - BitDefender.

 

AVG is good but the protection is really user-dependent.

 

IBK, will you confirm, was the "out-of-box" protection tested with Microsoft Security Essentials installed?

 

Forticlient did extremely well once again. It's a pity that I found it quite heavy in my system.
Avast continues to be probably the most reliable free solution, avg is also good, it may rely very much on user decisions about identity detections, but the messages are pretty clear and I don't think anyone could hit allow.

 

Sort: By value

PDF=
http://www.av-comparatives.org/dynamic-tests/

------------------
The 3A:
Avira Internet Security 2013
AVG Internet Security 2013
avast! Free Antivirus 8.0

 

Read the PDF, page 3

 

BD the best by far.

 

Page three does not actually mention whether it was installed or not. It is listed as a tested product, though, which is why I want the confirmation.

 

Wow. What is McAfee up to? Their numbers have really improved over the past few months.

 

AV-C should test Bitdefender Free as well.

 

IMHO, page 3 says more than this.

 

Where is Symantec?

It seems AV-C is choosing which tests to include Symantec, how unprofessional given that Symantec wanted to participate solely in dynamic tests!!!

Nice results for avast, wonder what is the user dependent part.

 

Things related to AutoSandbox (once it does not detect a malware) ?

 

The AV generates a pop-up which asking you what should it do.

 

Find me critical, find me like somebody who always break the party but what is written in the PDF and what AVC does is totally wrong. By default Windows Defender is in Windows 8 and is part of the OS, as well as other features. But Microsoft Security Essentials (although it is equivalent) is not part of the Microsoft OS and is not out of the box. For Windows 7 out of the box would be Windows Defender in Win7 + the Windows 7 security features.

Microsoft Security Essentials (although free and part of the Microsoft Security Client family) is not out-of-the-box for Windows XP/Vista/7 and does requires about 2 to 5 minutes manual interaction (downloading, installing, updating, etc).

Additionally, out of the box protection would require AVC or any other testing organization to test and count other Microsoft protection features such as Windows Firewall, Windows Defender, Smart Screen filter for IE, Protected mode, DEP, etc <these in Windows 7>.

Very unprofessional of AVC !

Additionally, they included Symantec just once trying to ruin its good reputation and to make it rejoin - now they skip it. This game did not work, AVC

 

BD free shouldn't be unmasked. AV-C need to eat too

 

Are AVs done at default settings?

 

Er....Symantec is included in the performance tests. It is just a matter of financial viability - dynamic tests are the most expensive (in terms of time and money) to test. No great "game" here.

I do think (and hope) that Symantec will probably be seen in more tests from AV-C.

Yes, I think the testing is performed with default settings, though vendors may ask to enable/disable/ignore some specific features (this is usually specified in the relevant reports).

 

What's the sample size?

 

Over on Fortinet's Forticlient forum, just earlier today, one person suggested turning off Windows Defender. A user was having serious performance issues, and in response, someone said, "Did you make sure Windows Defender is turned off? I have seen some serious slowness all over Win 7 if both are running at the same time."

In another thread here, King Grub linked to a new AV-Comparative performance report that showed some slowness in one or two tasks with Forticlient's previous version. Some AVs were a little lighter here and there, one (Kingsoft) was significantly heavier.

That said, I haven't seen significant slowness, even on a modestly powered PC (Core 2 Duo, 2.1 GHz). Maybe a classic case of YMMV.

 

Hmm..EAM was ahead in march and back in april,still questionable enough to me

The sample set is always a few 1000 whereas every month a AV see's just more than that.....you could check the last year's final results which has a lot of detailed explaination along with the sample size that pretty much wouldnt change

Avast did get better over the last month atleast it got better in the way the graph is actually sorted...otherwise so far it is doing well.

I dont see how testing with defaults is fair...in my opinion,there are many AV's that detect PUP's on defaults and testing avast with that off will be kind of unfair as there can be 50 in 100% chances of increase in blocked ratio with that turned on even if AV-C strictly tries to remove all the PUP's some malicious programs can be sorted differently by AV company.

btw,avast just released a new version fixing their sandbox logger so we can hope to see some differences in next AV-C test round.

Any word from Vlk about the missed samples in the test,last month the sandbox was catching on them what happened this time?

 

G-Data also scored 100% as far as i'm concerned. Even if it's user dependent, users usually get scared and block the detection anyway even if unsure.

I mean, if you treat user dependent detections as potential misses, then so you should every other detection that user can exclude. Because some will do even that...

 

Is it not financially viable to test MSE or Windows Defender ("Win 8")?

 

It's very fair cause all vendors are tested so and can influence their own recommended default settings.
But I think it's not fair if some products give the majority of users (which use default settings) less protection. Shame on them not on the testers

They test PROGRAMS and not USERS - so this treatment is quite ok, cause the programs don't give clear results and belong to users choice. And if they choose there is a potential for misses. btw. the user choice results are treatet as they are - no clear detection and no clear miss. (but we had this discussion several time.)

 

I dont see the point...avast doesnt give PUP on by default because it can result in detection of programs that arent malicious but can be used for bad purposes so how is that less protection,I am looking from the testing point of view: is it fair to test 1 product with something on and the other with same thing off!?

Another plainly stupid assumption in action to defend independent testing organizations and their methods of testing and attempt to attack a AV vendor,what's the deal are you guys being paid by AV test groups or what

I was just stating my opinion and point of view,if you dont like it ignore it,no need to stand in defense.