If you could turn a substring into any single repeated character, which character would be the cheapest to aim for? The best candidate is the one that already appears most frequently in it.

For a substring of length $L$ where the most frequent character appears $f$ times, the number of changes needed to make it uniform is $L - f$. When is this feasible within the limit $k$?

The answer is a single integer — a length. Could you determine whether a feasible substring of some length $L$ exists without examining every possible substring of that length?