Software Review Can't believe in signature test Kaspersky get outperformed by ClamAV

[Khushal]

Content source

https://youtu.be/Tw-WJksKYpo

 

[ForgottenSeer 114717]

Another Youtube tester that is not getting it right.

The AV engines submitted to VirusTotal are not the same as the ones in the products, which rely heavily upon cloud-based reputation file queries.

At any given point in time, Kaspersky does not detect tens of thousands of malware that other detection engines do. This is nothing new. So what is the point of this test video?

 

[harlan4096]

Also, at the beginning of the video, it shows a sample that Kaspersky can't detect in VT... it says last analysis 5 days ago , just stopped watching the video at that point...

 

[Parkinsond]

Another Youtube tester that is not getting it right.

The AV engines submitted to VirusTotal are not the same as the ones in the products, which rely heavily upon cloud-based reputation file queries.

At any given point in time, Kaspersky does not detect tens of thousands of malware that other detection engines do. This is nothing new. So what is the point of this test video?

The title of the video is direct and clear "signature test" which reflects offline protection; it did not mention online protection.
Regarding the field of offline protection, it is documented K, McAfee, TrendMicro, MSD, Panda, and "old" Norton were not top performers; ESET, Avast-AVG, Avira, F-secure, and GData can perform better offline in comparison.

 

[Parkinsond]

I have performed a similar test several times, and I may disagree with the reported results.
For me, K was at the top, followed by ESET, then Symantec and MSD, with Avast-AVG coming after, followed by B, and finally Avira; I did not bother checking the remaining AVs (was evaluating the free AVs in addition to ESET).

 

[Khushal]

The author of the video is in the forum, i expected him to shed more light on this test. That's the reason i posted this here.

 

[Parkinsond]

The author of the video is in the forum, i expected him to shed more light on this test. That's the reason i posted this here.

I can predict there will be a lot of "shedding" by several members for such a controversial video

 

[Khushal]

I can predict there will be a lot of "shedding" by several members for such a controversial video

yeah he seems to have a bias against Kaspersky. Maybe his own solution can't beat Kaspersky that's why.

 

[Parkinsond]

yeah he seems to have a bias against Kaspersky. Maybe his own solution can't beat Kaspersky that's why.

I decided to stop using K after declaring the start of deprecating its free tier, even if mine was still working, but K is and will remain a top product.

 

[Khushal]

Another Youtube tester that is not getting it right.

The AV engines submitted to VirusTotal are not the same as the ones in the products, which rely heavily upon cloud-based reputation file queries.

At any given point in time, Kaspersky does not detect tens of thousands of malware that other detection engines do. This is nothing new. So what is the point of this test video?

I hope my Youtube testing starting Q1 2026 will not dissapoint u. It will be the most unbiased independent cross checked AV tests on YT.

 

[Parkinsond]

I hope my Youtube testing starting Q1 2026 will not dissapoint u. It will be the most unbiased independent cross checked AV tests on YT.

Congratulations I am confident it will be great.

 

[ForgottenSeer 114717]

How many people know that the detection engines that vendors supply to VirusTotal do not work the same as the ones that they use in their products?

How many people know that due to multiple variables even "top performing" AV detection engines on VirusTotal can miss tens of thousands of legitimate malware while much lower performing ones do detect them?

How many people know that scan engines automatically copy other vendors' detections as a matter of routine operation?

How many people know that most security software is heavily dependent upon cloud-based queries and that VirusTotal scan engines will not detect malicious files that the product would otherwise because of cloud queries?

Without fully understanding these facts, how can a "tester" infer that one VirusTotal scan engine from vendor A is failing when scan engine from vendor B is succeeding?

 

[ForgottenSeer 114717]

I hope my Youtube testing starting Q1 2026 will not dissapoint u. It will be the most unbiased independent cross checked AV tests on YT.

Only Leo from The PC Security Channel gets Youtube testing right and remains 100% unbiased.

 

[Parkinsond]

Only Leo from The PC Security Channel gets Youtube testing right and remains 100% unbiased

I can remember the last couple of videos by Leo were criticized for methodology by MT experts; you do not have a consensus and will not have one; as you have said before: just pick the AV you like.

 

[TuxTalk]

Ok let the games begin......

 

[ForgottenSeer 114717]

I can remember the last couple of videos by Leo were criticized for methodology by MT experts; you do not have a consensus and will not have one; as you have said before: just pick the AV you like.

There are many Leo detractors here at MT. There are many pro-Youtube tester supporters here at MT. Hence my statement that Leo is the best and only unbiased Youtube tester. Get it?

I really do not care if Leo gets it right or not. When I see others bashing Leo, I will defend his position - just because I am smart, experienced, and knowledgeable enough to see what he is getting at - even though many times his "demonstrations" are not articulated very well and most viewers are not native English speakers and therefore do not fully understand what he is stating. His points are valid. I have never seen a single instance where Leo made an unfounded or illegitimate claim.

 

[Parkinsond]

I will defend his position - just because I am smart, experienced, and knowledgeable

Indeed you are smart, but some of those criticizing Leo are smart too; which smart should I follow?

and most viewers are not native English speakers and therefore do not fully understand what he is stating

I am not native English speaker too, but I can understand what Leo means and I enjory his videos.

 

[Trident]

Only Leo from The PC Security Channel gets Youtube testing right and remains 100% unbiased.

He doesn’t…
And he is not unbiased, he is on the principle “I like the one that pays”.

He was a Sophos lover for very long time.

He criticised products like Norton for slow removal of threats with very little knowledge why it was slow.

I can keep going.

He is certainly trying his best but his best is simply not up to the standard.

Also, VirusTotal can’t be used even as signature test.
And what is one trying to prove by testing signatures in 2025?

 

[Parkinsond]

And what is one trying to prove by testing signatures in 2025?

Trying to prove some products can be manually updated and used on air-gaped PCs used for gaming or productivity (rednering, multimedia editing) and some other products cannot as they rely on internet connectivity to provide the promised protection.

 

[Trident]

Trying to prove some products can be manually updated and used on air-gaped PCs used for gaming or productivity (rednering, multimedia editing) and some other products cannot as they rely on internet connectivity to provide the promised protection.

And when you manually update today, what happens tomorrow, by which time Bitdefender has released already more than 15 updates?

The statement that products like Avast and Sophos with 250 mb databases or products like BD with 2x 500 mb can be used offline is misleading.

It’s like saying that WhatsApp for desktop is useful without connection.
Yes, you’ll see some messages. But you cannot use it to its full potential.