Discussion in 'other anti-virus software' started by pcslinfo, Aug 23, 2010.
I second that
So, you believe that AVG suddenly became so Strong that it now exceeds G DATA
Meanwhile, Avira AntiVir suddenly became so Weak that Rising went ahead of it
The September AV-Comparatives.org 'On-Demand' Test will very soon clear the situation
and show us whether the PCSL Total Protection Tests are to be trusted...
not talking abt resources .. its just that their test results were always so poor that it was tough to imagine it competing in main road like today
well, i completely agree with what u say , i m not saying that these results perfectly reflect the av strength but note that its 3 month scenario detection .
.. look my above reply for what i meant
also looking forward to av-comparatives report
Thank you for your test!
I have some questions:
1) In your methodology you write: "Based on the monitoring data collected by
our own crawlers and by security vendors’ telemetry systems, we divided 2000
prevalent malware samples according to their level of prevalence into the following four groups:
Category A – Very High Prevalence
Category B – High Prevalence
Category C – Medium Prevalence
Category D – Low Prevalence."
Which parameters do you take into consideration when you rank files between four groups ? I don't see this information in the document and this information is very important for the chosen method.
2) In your methodology you write: "By using the analytical hierarchy process
method, we determine the weighting of any category, so that the full score
a security product can achieve from detecting malware in each category will be as follows:
Category A: 52.6
Category B: 25.3
Category C: 16.3
Category D: 5.8."
How do you calculate these coefficients ?
PCSL Total Protection Test (July 2010 Report)
Compare the above Results with
the Results of the last AV-Comparatives Test .
Needless to write more...
It all depends on the malware samples! It's so easy ...
The 900000 Sample of AV-Comparatives looks, by far, more Representative than the 2000 sample of PCSL.
BTW, it does Not only Depends on the Malware Sample.
It depends on the Testing Methodology, too.
See the #55 Post.
I agree, I don't really like the categories
If the product simply prompts the user for a decision, it sounds like the test counts this as a valid detection, which I think is BS.
If this is true, then just based on this alone, I think this test is bogus.
Both great questions.
Separate names with a comma.