I've noticed the narrative from alternative products that state they are less than HB in ban waves. Now what that imho means is that its either likely due to CR specifics only and therefore less pattern of behaviour / footprint left behind and/or HB is popular therefore its patterns are studied more. Either way, i don't rate anyone particular product higher or lower than the other and for me really comes down to a few points:
Popular can be healthy. If more people use HB thefore when a ban wave hits the alarm bells ring louder. Alongside this you also have alot more vibrant community pumping content into the product(s) - something i've noticed the others lack.
The downside is imho the bans are likely a result of this forum in many ways, the idea you post log files in an open forum for me is insecure. I'd prefer that behaviour stop.
So the chances of a ban really come back to your signs of abuse. If you leave your toons follow the same pathways over and over and over with exact X, Y, Z HotSpot coords, well, what are the chances a human does that? If you use HB inside Raids/Dungeons and your toon reacts instantly to the movement of a Boss before others ..again...once sure...twice...maybe... everytime...are you freaking kidding me? If you're farming mobs and you follow the same continuous loop ...again... human or bot?
Point is imho to downsize your noticeable behaviours would require you to layer in imperfection on your bot and make it clumsy as much as you can to throw off the data. Lastly, i'd rate the detection to be random sampled unless someone in-game flags you for cheating in which case that puts you into a check queue, which in turn cross-references all the previous points and outcomes a score. If score high, ban.
That's where i think the chances live anyway.