@tekknolagi Hmm, this sounds like a good situation for suffix arrays, augmented with longest-common-prefix arrays. There are algorithms to construct both in linear time, and then you can find the most frequently repeated substring of any given length in linear time as well. I think you can also quickly find the substring which maximizes the total length of that substring's repetitions, although I don't know whether that's also doable in linear time. And frankly I just think suffix arrays are neat and want to see somebody use them for something 😂