ReadonlycandidatesReadonlyexcludeReadonlyminimaxStatic ReadonlyisStatic ReadonlyneedsStatic ReadonlyneedsStatic ReadonlyVariantsRanking built by repeated wins instead of the method's own full ranking: run the method, record the winning tier, then re-run it restricted to the remaining candidates for the next place, and so on.
Differs from ranking() whenever the method's full ranking disagrees
with how it ranks subsets — e.g. instant runoff orders losers by
elimination time, while iterating re-elects a winner at every place.
Result of the vote. The first item lists the winners of the vote.
For example this ranking means that Bear wins, Sheep is second and Lion third
[ [ 'Bear' ], [ 'Sheep' ], [ 'Lion' ] ]
Requested in https://github.com/lzear/votes/issues/76
The Minimax-TD process:
Example 1:
A, B, and C are the members of the Smith set.
Look for the lowest 'worst pairwise defeat' in all of the comparisons (not just within the smith set):
B has the smallest 'worst pairwise defeat', therefore B is the winner.
Example 2:
A and B are the members of the Smith set. Both of them beat every other candidate but they tied each other.
Look for the lowest 'worst pairwise defeat', and since there are no defeats, look for the lowest 'worst pairwise victory' (not just within the smith set):
A has the lowest "worst pairwise victory" against all the candidates, therefore A is the winner.