To explain it in the most simplest terms....

"encryption" is a method of hiding something in such a way that YOU know how to "unhide" it and THEY can't work out how you did it.

If you try to use repeating patterns (like most of the examples given here) where I put the same letter into the same position and get the same answer... then I can BREAK the encryption by feeding strings into it that I know and using the output to "reverse-engineer" the tequnique you used.

The whole reason 128-bit encryption works so well is NOT because its good... but because it knows it would take forever to actually break the code.

I would suggest instead using some form of randomized positional encryption with a decoding array. This way parts of your encrypted output are actually "indicators" to the source reading it how to decypher the randomization... IN this way you could "encrypt" the same string twice and never have both equal to the same value and STILL maintain encryption.

It can still be broken but it takes people a damn sight longer to work out what you are doing than what is feasible... and if they are able to crack it then they bloody well deserve to get what they want because they are that good