When an account has more than a little history here, we tell people that we've banned them, and why:

https://hn.algolia.com/?sort=byDate&dateRange=all&type=comme...

https://hn.algolia.com/?sort=byDate&dateRange=all&type=comme...

Brand new accounts posting trollish or flamebaity stuff get treated differently. Software filters some of those, and we shadowban some of them, especially if there's evidence that this is someone we've banned many times before. The converse is also true: if software has filtered out a new account that's not posting trollishly or flamebaitily, we restore their posts and mark the account legit so it won't happen again.

I think this is a reasonable balance between transparency and defending the site against abuse. If we tried to give every banned account the same high-effort attention that we give established users, we'd do nothing else all day and still not get through them all. That would just be a new vector for people to DoS the moderators. A small number of abusive users can create a large number of disruptions. Shadowbanning is an appropriate tool for those cases.

Edit: there is also a significant amount of spam, and if we told spammers we were banning them, they would spam us with emails demanding attention, asking why, and telling us how high-quality their articles really are. Actually they do this a lot already, and it's a pain.

I'm not positive on that statement. But I have seen political bias and ranking bias quite often on HN.