@RileyHunter is correct, since I haven't finished the kata there are no partial spoilers for me. Thank you for breaking up the comment for me, that's very helpful information.

While I appreciate the rapid reply, it's very frustrating to have the comment be hidden as if I wanted to "spoil" I could just see the solutions and immediately understand what I'm missing xD

I'm confused. This looks like a TSP problem, however the javascript tests take > 6 seconds to run when all the function does is return immediately. That implies in order to successfully solve the kata I would need an O(n) efficiency approximation within 1e-9 accuracy of a TSP problem and there is no such algorithm. Not sure what I'm doing wrong.

nevermind... i give up... on everything

I know how you feel. I have an O(c) solution for all but one case, but I can't figure out the last step and feel I might have gone down the wrong path. After a week of effort it feels rough.

Linear time is too slow. Needs to be faster than that, and much faster than quadratic time (O(m*n)).

I don't understand why this would matter, mind explaining?

Yeah, php has a vulernability on codewars which makes any test without rng cases exploitable.
But there doesn't seem to be any activity on the platform from it's creators.

Edge cases such as ':' aren't addressed or tested for. Nor are there tests for ' '.

Wording is ambiguous.

Inefficient storage method, and you shouldn't build tests around them using your specific data structure "houses" unless you are going to provide it for them and make it mandatory in the kata description.

The kata doesn't specify that case should be ignored, but there isn't a test for palindromes with the same letters but different cases.