Reply to thread

Message

<blockquote data-quote="Andy Ful" data-source="post: 1123463" data-attributes="member: 32260">The AV-Test example.AV-Test&nbsp; Consumer Real-World tests 2023-2024.Avast + Bitdefender + F-Secure + Kaspersky missed 8 malware samples from a total of 3354 samples.The average infection rate of the top AVs:&nbsp;r = 1/4 * 8/3554 ~ 1/1800[ATTACH=full]288199[/ATTACH]Threshold criteria:p(t) = 0.05 for t = 1/1800Threshold table&nbsp;total ( n ) ........missed ( t )100&nbsp; &nbsp; .....................1700&nbsp; &nbsp; .....................21600&nbsp; ...................3The Real-World tests included 200-400 malware samples, so any result with more than 1 missed sample should cause the score to be 5.5 or less, exactly as in the AV-Test awards.These results do not prove that the method used by AV-Test is the same as described in this thread.Edit.Corrected a typo error in the formula for p(x).</blockquote>

[QUOTE="Andy Ful, post: 1123463, member: 32260"] The AV-Test example. AV-Test Consumer Real-World tests 2023-2024. Avast + Bitdefender + F-Secure + Kaspersky missed 8 malware samples from a total of 3354 samples. The average infection rate of the top AVs: r = 1/4 * 8/3554 ~ 1/1800 [ATTACH type="full" width="262px" alt="1745315464614.png"]288199[/ATTACH] Threshold criteria: p(t) = 0.05 for t = 1/1800 Threshold table total ( n ) ........missed ( t ) 100 .....................1 700 .....................2 1600 ...................3 The Real-World tests included 200-400 malware samples, so any result with more than 1 missed sample should cause the score to be 5.5 or less, exactly as in the AV-Test awards. These results do not prove that the method used by AV-Test is the same as described in this thread. Edit. Corrected a typo error in the formula for p(x). [/QUOTE]

Verification