Given a competition where each person is judged only by their other competitors, what is a fair way to minimise bias or foul play, and encourage accurate ratings? For example:
>one competitor rating everyone else zero
>one competitor rating fairly until they feel at risk of losing, then rating subsequent competitors zero
>good faith ratings still varying in how they use the scale
>singling out one opponent for an unfairly bad mark

A simple adjustment would be to award each competitor an average of what the scored everyone else. This would correct for constant bias, but not for focussed foul play. You'd need an additional way to weight ratings compared to the group.