• lacking sample tests.
• assertion messages aren't really helpful. => show one possible output to the user or at the very least give the minimum length expected

I didn not read the name of the Kata (usually it is not necessary), so did not understand that I have to to find bigrams only. I was looking for all possible сombinations... :( It was sad. Maybe you could add one more example ('abc'?) to description to make it clearer?

you can do it... keep improving your brute force solution until you stumble upon it

Should probably re-rank it to 4-5 kyu (in view of the 6x6 and 7x7 katas). What do you say?

And I passed the C test with essentially the same code...
Hopefully someone can check the C++ random test generator

let's do the 6x6 then ;)

I cannot tell more than: it's about geometry. And without loops.

Can you please give me more of a hint as to how to do it? I still can't figure out how to fix it. If you can't, I guess I could finally give up and look at the solutions. But it just bugs me that I am so close to the solution but not quite! (Maybe this is the wrong place for this)

Fixed. Thanks ;-)
And sorry for late(I haven't seen this issue report before today)

seek for something without any kind of loop (O(1) solution)

`

An algorith for n-gram generation, beautiful!

Then it'd be too obvious though ;-)

I think it's fine as is.