The best heuristics are those the candidate will never prepare or even think about. I don't think it's difficult to find one. What happens though is when it becomes popular and everyone starts doing it, then everyone starts preparing for it and then looses most of it's predictive ability.

Filtering in a sense is an arms race between finding new novel heuristics and ambitious careerists seeking to game those very heuristics. In the past it was the college degree, then extracurriculars, then leetcode, might even be these games in the future.