AV Comparatives - August 2007 report

Has anybody a clue when latest report will be out?

One thing I've never understood completely (no news there ) is why false positives cause AV to prevent a tool from a higher class like getting only standard certification!

I'd rather get 99.9 % of the viruses and a few quarantined FP's to deal with than 95% of them and 0 FP's?

But what do you guys think?

 

I think it will be out the 1st of September.
I won't be switching AV's based on it. I'll just stick to what works for me.

 

Re: AV Compartives is Almost Due

would you though, even a difference of 90 > 100 is only 1 virus missed per 10, and how many viruses do you actually get?

you should not forget these tests use hundreds of thousands of samples.

i would be more happy if drweb kept its 90 or there abouts, and lowered its fp rate, but thats just me.

 

Re: AV Compartives is Almost Due

I think IBK said the results could be a bit delayed.

 

Re: AV Compartives is Almost Due

nah it will be on time.

maybe even early, wink wink lol

 

Re: AV Compartives is Almost Due

I prefer having low FPs; it may not be that difficult to submit them, but Id rather not deal with it. And as NOD32 has proven, you can have both great detection and low FPs.

Also arent FPs only important in teh Retrospective/Proactive test?

 

Re: AV Compartives is Almost Due

you will not be able to draw a conclusion based on the outcome, as they will score pretty much equal.

 

Re: AV Compartives is Almost Due

Now thats a pseudo-cryptic response if I have ever read one!

Pretty much equal, eh? Then let the fanboy squabbling commence!

 

Re: AV Compartives is Almost Due

OMG i can't wait to switch my AV and my avatar for the new N°1 !!

ohh wait , i don't use an avatar anyway

 

Re: AV Compartives is Almost Due

Shh!!! AV X scored 1% more than AV Y!!!!

 

Re: AV Compartives is Almost Due

Switching AV programs because of any particular "test" result is like taking a very small sample size and making a general conclusion. The only way to logically look at the efficacy (or inefficacy) of any AV program is long term results.

I am always amused at the almost knee-jerk response to say a test like AV-Comparatives. Look, if NOD32 is down a little on a % basis, so what? Is it a meaningful (statistically significant) event or just a deviation from the mean?

My view is that you must assess the value of any AV program by long term results. A few months (good or bad) is an indication of nothing but a few months of good or bad.

 

Re: AV Compartives is Almost Due

I don't keep one for a few months if it is "bad" on my system. I sure would not keep it if I were getting infections either. But, I have never had an infection.

Best,
Jerry

 

Re: AV Compartives is Almost Due

My point was a few months of results are irrelevant. A few months does not make a trend. And, I have never had an infection either. I have stated in this forum before that people with a lot or virus problems need to look in the mirror for the most likely cause.

Obviously, any AV program should not be on your system for 5 minutes if it is not compatible-and some AVs are not. For me, Bit Defender and AVK Internet Security (AVK AV only is ok) do not play well.

 

Re: AV Compartives is Almost Due

Hi CSJ: As before my comments in red to stay OT!

 

Re: AV Compartives is Almost Due

TY, I can wait no rush!

 

Re: AV Comparatives is Almost Due

because fp rate does not give a reliable av.

fortinet with its 1000+ fp's is not realiable at all.

it gave me about 60 extra infections and this is compared to drweb *lol*

unrealiable av's that give 'too many' fp's should be knocked down.

i dont know what the level is for an av to get knocked down, but maybe it should be more than it is now i think.

 

Re: AV Comparatives is Almost Due

at VB you do not get VB100 if 1 FP occurs.
as we send all occured FPs to the vendors to fix them and considering that the grow rate of our clean set is not high, we are (currently) still quite generous.
The august report is not delayed, it will be released on 1st September.

 

Re: AV Comparatives is Almost Due

Hmm, still waiting for AV's rationale!

 

Re: AV Comparatives is Almost Due

ok vb100 do infact do that, but what size is their test set?

 

Re: AV Comparatives is Almost Due

Hi IBK:

I am glad your report is on time! Well done as usual!

What does "at VB you do not get VB100 if 1 FP occurs" mean? I'm to ask?

It's good you send the FP's to the vendor's so they at least have a chance to reduce FP's. Is it correct to interpret the slow growth rate of the clean set as meaning the vendors are NOT repairing any FP's? or are they just slow!
Maybe I should look to see which are most active in repairing FP's!

I still don't quite know why an AV with a FP is so bad if it just gets set aside and can be easily put back and maybe marked to ignore? Or is it more just an irritation issue? My view is different than some other posters here ( that's okay)

If AV 1 could get 100% of all malware and have 12 FP's while doing that would that not be better than AV 2 getting 99% and 0 FP's. My premise being that the 1% of viruses missed COULD be fatal to my PC?

None of this is personal so if my logic/thinking is wrong in this example I'm interested in any poster's counter argument! Not a statement saying I'm wrong but the rationale on Why this position is flawed.

If IBK can't comment on this due to the testing rules that's okay!

 

Re: AV Compartives is Almost Due

Hi WS Fuser:

If we can get both that is good!

 

Re: AV Comparatives is Almost Due

no, it only means that lately we do not add hundreds of GB's of new clean files to the clean set.
the FP's that are encountered in the test are immediatly send to the vendors and they fix them all within hours/days.
An example: some users (not so experienced as e.g. you) which AV did delete or quarantine a file (a false positive) belonging to a program which they use and which after that does no longer work or act strange, call a computer repair company and ask for help saying that recently they had a virus on their PC. That company then grab the PC for some days/weeks and then asks e.g. 65€ (that's what it usually costs here in Austria) for "repairing" the PC. Imagine this happens in a company, even if a small one. Lots of working hours and money gets lost due "just" a false positive. A false positive can cause as much trouble as a real infection.

 

Re: AV Comparatives is Almost Due

Very good, I grasp the reasoning now, so then I can have my theoretical 100% AV and 12 FP's because I can restore the program as a semi experienced user! Okay so be it.

The poor inexperienced user gets 95% detection an zero FP's to avoid repair bills he shouldn't have!

But meanwhile this same poor guy with no FP repair bills has the risk of his AV missing a really bad virus that destroys his whole set up! Now that IMHO will be an even bigger repair bill!

But never mind, to each his own!

 

Re: AV Comparatives is Almost Due

It isn't even about the average guy. Some people like to say that corporates absolutely love paranoid protection, but a certain Korean AV vendor used to tell me otherwise and stressed the importance of having lower false positives on a corporate environment as well as home user environment in order to maximize functionality and productivity. In the end its a compromise, and the best compromise has to always be made. And in many cases users do not like annoying pop-ups getting in their face every second, sometimes they get alarmed by such pop-ups, sometimes it causes problems like IBK mentioned. 65 Euros for one PC, how much for an entire corporate network brought down by an FP?

For all the above reasons, such users are going to drop the product and go with something else. Of course, different people's needs differ.

 

Re: AV Comparatives is Almost Due

I have to agree with VB100, one FP is too much. In an enterprise setting one FP is a disaster if it happens to be in a file that has been imaged onto hundreds or even thousands of machines. Or how about an FP on an OS component that disables the machine. This happened a few years back wih F-Prot and more recently with Symantec on Chinese language Windows.

Its a gift if a bunch of AV's all score high. That allows users to make a choice on important stuff like lack of system slowdowns and having an understandable interface.