You can't invent tests

But I can, because it's a valid test case. I just wish edge cases like that existed (if this kata wasn't ancient, they would).

Of course it will fail if you start making up random stuff, all of them will.

No, they won't. Only flawed solutions similar to yours will fail. The description and constraints are pretty clear, so there is nothing to "make up" to make this solution fail.

It's not random stuff though. It could've been in the tests as it's a valid case.

I agree, this is not the most efficient code but it's still clever.

We should avoid this at all cost

It's still O(n), it's not that big of a deal.

explain why and when does it fail

"Wrong" doesn't mean "unreadable" or "not performant", it simply means incorrect. Take an example like ['n', 'n', 'n', 'n', 'e', 'e', ...] (fill with 'nnss', for example): your code treats the first section as if you reached the starting point. Clearly that's wrong.

where n is always 10 and thus it's O(1)

meanwhile, your own solution is outright wrong x)

Awful

Can't understand why some people voted this is best practice when it's the exact opposite. Unreadable code with a lot of bloating comments.

This is fun to do and very clever, but such unreadable code.