Changes to Activision ShadowBan Policy?

So I've been seeing threads reporting HWID bans and how people are changing their serials and installing a fresh copy of windows following a ban.

I got banned in the last wave just for using a russian account that I got to get MW for cheap. I then made a legit US account and it's shadowbanned but NOT getting permabanned. I saw the thread that said shadowbans typically last 7-14 days however there are rarely people that receive HWID permabans which results in every future account they create not only getting high ping lobbies but shortly after getting permabanned. I just want to ensure this is still how the shadowban paradigm works.

It's takes 2-3 accounts to get a perma and depending on the detection it could lead to a perma. When Engineowning got detected it was a straight perma even if you hadn't had an account banned previously. To Get rid of a Hwid Shadow ban you'd need to reformate you drives which will result in a serial change for those drives

It does not take 2-3 accounts. Nowadays you get perm bans really easy. Since they are haunting all spoofers like crazy it’s pretty easy to get a perm ban.
If you are shadow don’t play the Game for like 14-20 days. Otherwise use a spoofer for 1 month and hope for no detection ^^

I've been perm banned on multiple accounts, still not HWID banned. Maybe I'm just lucky and it's probably easier to get perm banned at the moment. Like @[Only registered and activated users can see links. Click Here To Register...] said above, they're targeting spoofers to catch repeat offenders, it's a real risk right now.

Quote:

Originally Posted by GrawPoint I've been perm banned on multiple accounts, still not HWID banned. Maybe I'm just lucky and it's probably easier to get perm banned at the moment. Like @[Only registered and activated users can see links. Click Here To Register...] said above, they're targeting spoofers to catch repeat offenders, it's a real risk right now.

Experiences like yours makes me thing they use machine learnings cause it smells non-deterministic; basically they're going off of statistics which is at the heart of machine learning.