Has there been an "arms race" of sorts with programming contests, as knowledge percolates and harder categories of problems need to be chosen? Since these days dynamic programming is basically table stakes for ICPC-type problems.
I think it's because access to knowledge has become a lot easier, so contestants know more; plus, the competitions themselves have become more organized.
For example, last I checked, the IOI had a syllabus outlining the various algorithms contestants needed to know. During my time, it was more of a wild west.