Hacker News Reader

Comment by rm999 | Hacker News Reader

Am I looking at the wrong code? I see

  SORT (ups + downs) / max(1, ABS(ups - downs))

which would give a way higher score to the 1000/999 case than 2/2 - 1999 vs 4.

Hmm, perhaps I slightly misremember, but even that algorithm still scores +2/-2 as more controversial (result=4) than +1000/-500 (result=3).

And that's right. It's a controversy measure without taking in consideration popularity. That can be applied to the final score separately...