February On-Demand Predictions

I will predict the results including DOS viruses/malware. GData AVK, Avira, F-Secure, Kaspersky, Nod32, TrustPort, Norton, BitDefender, Norman, Avast, McAfee, F-Prot, AVG, Dr.Web, Microsoft

 

i think kaspersky will be either the same or better than f-secure.
because the new heuristic engine should be out by then.
lodore

 

please note that in 2007 the DOS category will not be counted anymore.
fortinet and escan will be also included.

 

This will depend on the deadline Andreas has for the feb on-demand, but i do not think it is possible, i think we need to wait until the next on-demand.

 

i think he should not tell the vendors when the test is done,he shouldnt even ask for their permission to test such software, if he does .. how is it valid when some companys, will just add loads of sigs before the test, how does this make it a valid test?

 

What really I thought the DOS category was counted separately

 

I guess ill change my predictions to Avira, GData(heard they switched engines), F-Secure, Kaspersky, E-Scan( I heard it is like Kaspersky, not sure if this is true), TrustPort, Nod32, Fortinet, Norton, bitDefender, Norman, F-Prot, Avast, McAfee, AVG, Dr.Web, Microsoft

 

i await the day.... the companys dont know when the test is, only then will people see a valid result, we know for fact... some companys, not naming any names add tons of sigs before the test date, how does this make a valid test? i really dont understand it, why go to all this work for invalid results,really?

 

At the very least, the tests do work a great deal towards getting the vendors to add signatures, thus eventually increasing end user protection in the long run

 

i do understand that, i really do.

but if its a test for detection, surely without them knowing, so no cheating is going on, would be the best to see who performs and who doesnt.

then, the companys scan add the sigs to try and improve before the next test, this will show how they can keep up, show their heurisitcs helping with detection.

i think, if they know when the test is, its just not valid is it? ... because we all know, some add loads of sigs just before the test.

 

not invalid at all - think again . if you observe regularly when companies add large quantities of samples you will notice that it is always at the same time (even before av-comparatives existed this was the case). This have not to do with tests, it has to do e.g. with when they get the samples shared from other companies and the time they need to add them. additionally, there are some months in the year where companies release new engines etc. which will may look like they added a lot in short time (which is not the case). E.g. the last times when e.g. ESET added a lot of samples before the test, those samples they added where not between those that they missed in last test, but where other samples. The ones they missed in last test where added weeks before (and I think I already explained why it makes sense to release one big signature update instead of smaller ones - it simplifies the QA) - if another one adds tons of samples within 1 day after the last test, what quality do you expect from that additions?. The tests show how much the various AV detect. Usually after each test I get mails saying "some weeks after your test AV xy released big update, it would have scored differently with it" - sounds like a quite limited view of the things .
It does not count when a sample is added, as long as it gets added. Even if they know when more or less the test will start, seems like some companies do not even try to improve, while others do and contribute to better protection for their users.

 

IBK if this is infact the case, simple... dont tell them when the test is.

 

You can already look at here and there!

Best regards,
Firefighter!

PS. Almost forgot this this in post 1, my fault.

 

yes i know of the pcwelt, cheers firefighter

its an interesting test to read, not much difference between alot of them, quite of number of them in the 90% +

was just wondering if av-comparatives didnt tell them when it is, i think it would be a more valid test.

 

Why are certain AVs' not tested such as CA?

 

read the faq on the website (or read the second and third sentence on the main site)
e.g. only AV's with high detection rates are included in the test.

lets stay on topic.

 

To IBK: How about reporting the detection rates for samples received during the last 6 months.
Some antivirus are good for old samples but for new malware...

 

I do it the other way; in the report you can see how much of the missed of last test were added. its all in the report, you just have to "study" it

 

Yes, I noticed that section on the reports. Pretty good work on that.

 

i dont think on-demand will use heuristic engine so kaspersky result wont be affected by new heuristic engine

 

On-demand tests do use heuristic engines. If Avira didnt use its heuristic engine it would score significantly lower

 

@IBK - I would have asked this at the AV-Comparatives forums but they are down.

What/When, if any, are the "Special" tests for our beloved under dogs Rising, VBA32, etc.

 

I expect ranking to be about the same, but AVG and F-Prot will move up into the Advanced category.

Jerry

 

February/March I think. If there will be enough time and if VBA32 scores over 85%, it is possible that VBA32 is in the main group - will try.
Forum is down because I will be soon away for 1 week (back to Italy) and therefore not be able to delete in realtime P) all spam messages from the forum.

 

hello,

i hope you will keep them if they reach to minimum. thank you.