The kata is completely wrong: For strings with even length there are more than one possible set of triplets (e.g abcdef can be both abc, cde, def and abc, bcd, def), and yet your test only accepts exactly one configuration (where the bigger overlap is always at the end).

Input value is string in sample tests, and list of single characters in the actual tests.

All good now! :)

Thank you! :)

Green light!

Ok, there it is.
Thanks.

Thanks, but it would be nice to have this link right in the kata description.

http://www.codewars.com/kata/recover-a-secret-string-from-random-triplets

Can't find "Recover a secret string from random triplets" kata. Could you please provide a link to it in the description?

There it is. "Random triplets test generator". Still beta.
Not so easy.

That's definitively faster: O(1)
(More math than informatics, though)