I keep reporting accounts that call black people "nic gurr" to avoid being caught dropping actual N bombs and they are almost never banned.
I guess Daily Active User KPI is more important!
Edit - For anyone wondering about Trump's place on Twitter, he's not going anywhere. Twitter is extremely proud of his account and is using him in advertising in Japan:
http://www.businessinsider.com/twitter-ads-featuring-donald-...
This is a war you'll never win, though. People will keep inventing new combinations of characters as you ban the previously used ones.
Coded language like "states rights"[1][2] sill slips by a lot of people today.
[1] https://en.wikipedia.org/wiki/Dog-whistle_politics#United_St...
[1] https://www.theguardian.com/technology/2014/aug/21/twitter-f...
The word "if" was probably the key factor here. Since his threat was conditional, that most likely puts it into a different category than outright threats.
Twitter also enforces its policies very selectively, and seems to be getting away with it so far, but they are walking a tightrope. E.g. now violating the ToS is OK if it's "newsworthy" etc. Or how "Kill all X" will get you banned for some values of X but not others.