Pimpampum, I see you are around for more than a season, then please, do refresh your ban memories, and check the past Blizzard's ban records.
So far, in the last 20 years, Blizzard ALWAYS had banned all accounts, detected for exploitation in timely matter - no matter if it was botting/duping or random minor minor exploit.
So the belief that Blizzard do NOT want to detect some behavior or do NOT to ban if detected ... is just immature and not correlated with the real world.
No real reason to discuss this because....
Neither you nor me know a crap about it.
making a statement like you just did, " in the last 20 years, Blizzard always had banned all accounts detected for exploitation in timely matter" is wrong in every word, buddy.
There are people with accounts surviving every single wave until now, and having botted since 2010. Extremely rare, but true.
We don't know how Blizzard operates, we just speculate. Making statements about it is plain dumb. Plus, 20 years? In D2 world, that goes back to botting as far as I can personally remember, banning patterns had nothing to do with what seems to be the norm now. At all, and anyone that was botting back in the day will tell you so.
There has been a massive banwave recently. This banwave, that apparently was survived by a few people, was a one-time event thankfully. Considering this is Blizzard's common procedure, when it hasn't for the last 5 years, is shortsighted.
What we do know, is different penalties were applied in the past to different actions. So, they have their ways to identify an account measure of offense and they do act accordingly. They don't blindly ban anyone they detect, as you say.
They do ban when they want, who they want. At 1 point in the last 5 years they banned everybody they could detect; the rest of the time, they have been acting like I stated.
We are talking about a multimillion company here; thinking their procedures are easy to detect and determine is naive. Too many complex aspects come into play here, including suscription numbers, bot effect on the economy, closeness of an addon / expansion....so many things.
Another way of looking at this is the manual investigation. What Blizzard claims, is when they ban an account it is previosly checked by several Gms. So , they check in game behaviour for botting patterns. There is a huge degree of subjetivity in there; what GM 1 considers obvious botting, won't be what GM 2 considers it. Plus, There is a number of actions (like setting jump intervals, close player targetting and so on) that will make even more difficult to visually determine if one player is a bot.
Bottomline: while bot1 could have been considered a bot by gm1 and gm2 and then is banned in the next wave, bot2 could have been flagged by gm1 and spared by gm2, giving it a time to survive and keep botting until further investigation is done.
That's where the subjetivity of "Who they want" comes. It has nothing to do with personal preferences or friendships. Both bots could have been "behaviorally detected" at the same scan or moment in time, yet they won't be banned at the same time.