>I'd be interested in a 12-word IRV explanation.... I don't think that's >possible, either. Derive successive plurality eliminations from ranked ballots, eliminating one candidate each round.