The Russians, and Vietnamese are at the cutting edge of view botting in a way that can bypass the algorithm.
The majority of SMM panels are unable to offer this as a paid service; hence the majority having view drops after the update.
They are still using the 30+ second view methods, which no longer works without interactions as I explained.
There are some suppliers with the ability to bot views without drops, with steady delivery, in quantity.
However this supply is expensive, and my source isn't something I am going to reveal to you free of charge.
Consider this; say I supply you with the method; that does not benefit me whatsoever.
What it does do is overload my supplier, thus slowing down my own personal delivery. A gain for you, a loss for me.
I can tell you your current method of botting will not work after the update.
The algorithm doesn't weight subscribed views, versus unsubscribed views differently; that makes no difference.
Interaction weights the view, the more interactions the better.
With the Russians, and Vietnamese they work on enormous volume; scaled up to levels you cannot imagine.
They don't need to use proxy IP addresses, as they have tens of thousands of gmail addresses with YouTube accounts ready to view content.
When the algorithm sees views coming from thousands of channels that interact across thousands of videos; they register their views.
You need to appreciate YouTube is mainly automated now, the human view behaviour has now been machine learned and automated.
The updated algorithm looks for human like real view behaviour; such as a view, plus video like, plus comment like, or comment.
They already know the typical real view retention is under 2 minutes, so it doesn't matter if your bot views 30 seconds, 60, 90, 120, etc.
The time stamp where the bot likes the content does not weight the view.
The point is you need to be botting in massive volumes, with massive volume interactions.
The key is the large view volume, and large interaction volume looks organic to the algorithm.
Lumping on 1000 views fast, and then liking 1000 times after 30 seconds isn't looking real to the algorithm; hence your accounts being flagged over and over again.
I will give you something else to consider; the time frame of the views is important.
Lumping on 1000 views within the first hour doesn't look real to the algorithm.
It needs to ramp up steadily; nothing in 0-1 hour, 1-2 nothing, 4-6 it begins, 8-10 it gets heavy, 12+ and delivery peaks, etc.
If you lump on 1K in hour 1 that flags the views as botted as that isn't real behaviour.
Content moves up the rankings not in the first hour, it takes massive views, and massive interactions to move content up the rankings.
Hence views and interactions peaking after 12 hours not in hour 1 etc.
The Russians and Vietnamese aren't just lumping on 100K views in hour 1, with 100K interactions; the delivery is all phased steadily.
Once the content moves up the rankings then delivery heats up to incredible levels of delivery.
So you see its not as simple as having a bot that watches content for 30+ seconds then likes a video.
That needs to be scaled up into the thousands of accounts, and delivery needs to be phased in steadily with time, not lumped on in too short a time frame.
I will give you an example. I am currently using a source; it looks like nothing is happening 4-6 hours after a video upload.
Then 6-8 hours on it heats up with delivery increasing fast with time until the full delivery is done within 24 hours.
No drops whatsoever, the algorithm registers the views. The typical interaction rate is circa 30% of the view total.
I will give you one interaction that the algorithm currently weights which causes it to accept views.
Changing the resolution on a video; from 1080 to 1440 or 2160 HD. This is human like behaviour, changing a video from 1080 to 4K when watching.
Say you have 1000 views going on, and they change the video playback settings from 1080 every view is registered due to the real interaction.
The point is, the interactions need to be varied, not just 300 likes. The variety of interaction is what bypasses the algorithm on massive volume being lumped on.
10K views with 300 varied interactions looks real to the algorithm. 300 video likes in 0-1 hours doesn't.
