Such measures are pointless.
Someone finding obscure Unicode characters to bypass filters won't be stopped by banning of any specific character. Even if you ban every non-ASCII character, there's good old underscore. Serves same purpose, just isn't as pretty.
Unicode is constantly updated. It's incredibly tedious to keep up with it. Hint: there're already ...
Chinese characters are blocked in posts network-wide, with the exception of a few sites (including this one), because of persistent spam posts that use them.
The error message is intentionally kept nondescript to prevent spammers from catching on to the real cause why their posts are being blocked.
tl;dr: spammers prevent us from having nice things...
As a workaround you can still provide the original message, but in base64 encoding and instruct readers on how to decode it, like using Notepad++ plugin MIME tools. This will prevent the original text from being immediately visible though, so any search engine queries would fail to find it probably.
According to "Writing mathematical expressions" by Jukka K. Korpela, page 128:
for [unspecified] historical reasons [...] many fonts have [φ and ϕ] reversed, such as Calibri, Consolas, Lucida Sans Unicode, Palatino Linotype, and Symbola. This means that its generally very risky to use φ and ϕ in different meanings.
Therefore, SE requires a ...