David Feinstein was in good form last night, as he took us through his specific challenge: to help judges make more efficient use of their time, when caucusing to pick "best in show," other levels of reward.
The mind-boggling feature he shared last time, has since been replaced by a much simpler technique borrowed from the economists.
Projects are pairwise compared in a matrix with 10K resamplings of judges' scores for each one, a kind of play-off in alternative universes (where the combinatorics come in). Clever economists.