The main purpose of this site is to study "the science" of rating sports teams, especially college sports, where the schedules are highly unbalanced. This is the definition of a metric I use to compare rating systems, the Weighted Retrodictive Ranking Violation statistic.
Recall that a retrodictive ranking violation for rating R occurs when after the game has been played the winner has a worse rank than the loser. The Weighted RRV is larger depending upon how highly the rating has ranked the losing team and by how many scores the worse-ranked team won by:
| ||||
|
To calculate the Weighted RRRV for a ranking, just sum the violations for all games where the loser has a better rank than the winner and divide by the number of teams ranked by the rating.