Even a relatively simple cipher can offer some real protection in certain situations. I scribble private notes to myself in meetings in a slightly modified futhark. I know lots of people who can decipher it relatively well, given time, but I don’t know (or know of) anyone who can read it upside down at a glance–I’m always shocked when people are shocked that I can read (English) upside down or backwards at about half speed (which, being a bibliophile, is pretty fast). I always assume that others can read anything in their native language that they can see. Any system that requires even minimal analysis requires more time than a passing glance will allow.

As for the contest, you have to consider requirements on length and content, especially length. Otherwise, I would submit “XDF OPQ” and even tell you that it’s a simple substitution cypher in English. Decipherment is impossible, because it could be any pair of three letter words without any letters in common, and I don’t even have to decide what they are until all the “solving” is done and I choose something that makes all the guesses wrong.

Limiting content in some way also makes decoding the Cryptorthography samples more feasible. It’s so much harder to decode when you have no idea what the subject matter even could be.

Ha! I just had (what I think is) a great idea. The sample of your Cryptorthography for the contest should be a snippet that *explains* your Cryptorthography system! That makes solutions somewhat self-verifying, limits content in a meaningful but not overly restrictive way, and makes sure every entrant has something relevant to say on the subject. Decoded samples should explicitly exclude any non-Latin symbols or characters (it’s not fair, in a contest, to let use-vs-mention complicate the decoding even more!) The length of the unencoded original text should be some fixed length range: 50-150 words?

Thanks for the nice mental romp..