>>419
I'd rather not say, because it would be personally identifying, sorry. But you just find one, start from easy problems and work up to harder problems.

Note that the all-stars of this style of coding are shitting out low quality code because they have to get the answers fast; while you can learn very clever things from analyzing their code, it's not a way to write code that's supposed to be used for real (i.e. someone will have to read it someday).

A lot of the easier problems are variants on things you'd find in any algorithms textbook. A lot of the harder problems are actually too hard for me, but they're usually like multiple layers of the obscure problems found in the doorstopper textbooks programmers only claim to have read to look cool but haven't actually.