Reply to thread

Message

<blockquote data-quote="Andy Ful" data-source="post: 968144" data-attributes="member: 32260">To the MT readers.Please be patient. My posts are directed not only to [USER=66256]@Local Host[/USER]. Many people do not realize how complex is reliable AV testing. It is not sufficient to throw many malware samples and count how many of them were detected. There are many other important things to do that are skipped in the video tests. The final one is a statistical analysis of how reliable are the results. This is a similar situation to the election forecast. Any company that makes such a forecast has to calculate a statistical error, which usually is about a few percent. There is always such an error because we still do not know the votes of most people. So, if one has 1000 malware samples and the testing methodology causes 5% statistical error, then the AV that is protected against all samples can have the same protection in the wild as the AV which failed on 50 samples.If one takes into consideration this factor, then magically almost all video tests are compatible with professional tests (video test wonder).<img src="data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7" class="smilie smilie--sprite smilie--sprite109" alt=":)" title="Smile&nbsp; &nbsp; :)" loading="lazy" data-shortname=":)" />The professional AV Labs do such analysis. It is visible in the final recommendations. So, if AV-Test gives the 6 points to 10 AVs, it means that the testing methodology cannot say which one is better (even if they missed a different number of malware samples).</blockquote>

[QUOTE="Andy Ful, post: 968144, member: 32260"] [B]To the MT readers.[/B] Please be patient. My posts are directed not only to [USER=66256]@Local Host[/USER]. Many people do not realize how complex is reliable AV testing. It is not sufficient to throw many malware samples and count how many of them were detected. There are many other important things to do that are skipped in the video tests. The final one is a statistical analysis of how reliable are the results. This is a similar situation to the election forecast. Any company that makes such a forecast has to calculate a statistical error, which usually is about a few percent. There is always such an error because we still do not know the votes of most people. [B]So, if one has 1000 malware samples and the testing methodology causes 5% statistical error, then the AV that is protected against all samples can have the same protection in the wild as the AV which failed on 50 samples. [COLOR=rgb(184, 49, 47)]If one takes into consideration this factor, then magically almost all video tests are compatible with professional tests (video test wonder).[/COLOR]:)[/B] The professional AV Labs do such analysis. It is visible in the final recommendations. So, if AV-Test gives the 6 points to 10 AVs, it means that the testing methodology cannot say which one is better (even if they missed a different number of malware samples). [/QUOTE]

Verification