Measuring things would be one thing (number of repositories, number of pull requests sent, number of changes committed, number of comments made). I think the concern comes with boiling down a bunch of (possibly unknown) measurements into a single number. Who is to say that the formula is the "right" one or the "best" one? And what happens when people start getting ranked in an absolute sense based on that score?

And if people start changing their behaviour because they're gaming the system (or is it the system that gamed them?), and that results in unintended consequences, do we (as a whole) come out ahead from where we started?