Data by Ken Overton
Original idea by Daniele Paserman
The weighting was borne of an attempt by Daniele Paserman and others to come up with a way of "weighting" goals scored to exaggeratedly under-value what were largely considered throwaways: goals piled onto a rout, or pride-salvaging but competitively useless goals in losing efforts.
Toward this effort, I determined to weight goals based on the margin at the time the goal was scored. The goal which puts a team ahead is fully valued -- there may be three go-ahead goals in a game, but they are all valued the same because at that moment they are of primary importance. The basic value of a goal is the inverse of the margin created by the goal. Thus, a goal to go ahead by 2 is worth 1/2 (0.5); to go ahead by 3 is worth 1/3 (.33); et cetera.
So what about tying goals, and goals with a negative margin?
To keep it as simple as possible, I simply mirrored the curve from the positive side around the weighting curve's center at 1; so that a tying goal (margin 0) has the same weight as margin 2. a picture really states it best :
