It is the mean of rank, which is a float obtained from rating, which is also a float.
( rank: float = log(rating: float / A) * C )
It has various bounds on it (30k-9d, except for pros who are higher), but for simplicity I didn’t attempt to include the code from the “normal” implementation, which also takes into account large rating deviation.
That question terrifies me - I don’t even know what that is, in the context of auto-handicap 