Look at *differences* between words rather than at the words themselves?

[radapox]

So whenever you can whip up an at least 500-word code, I can tell you what it compares to.

Oh dear.  Well... I'm not making any promises, but I'll let you know if I happen to find the time (and coding skills, I suppose) for that!

EDIT: On a more serious note: I think a complicating factor is that a lot depends on how you map the mutations to the letters of your source language. If, for instance, frequent letters happen to correspond to more "drastic" mutations (ones with few zeros, in my method), adjacent words would become less similar than if the reverse were the case. So I expect that one source language may yield entirely different TTR results depending on your mapping.