Ask HN: Are superstar programmers usually good at CS research also?

Not necessarily. Research requires a different combination of skills - 1) ability to quickly read a research paper to get the main gist, 2) ability to estimate if a contribution is incremental or novel, 3) writing a coherent research paper, 4) defining a good experimental setup etc.

The Superstar program is hard to identify, rather than easy. Alex St John stated this.