[EM] please read quick question about Condorcet

[Forest Simmons]

Another use that has been suggested is this:

First calculate all of the pairwise matrices based on distinguishing more
and more levels.

Use the finest level matrix to calculate the Smith set.

Eliminate the candidates not in the Smith set.

Use the next matrix to calculate the Smith set of the Smith set.

Use the next matrix to calculate the Smith set of the Smith set of the
Smith set.


Etc.


If at any point the Smith set is a singleton, then from then onward the
result won't change, so stop.

The last step is sure to have a singleton Smith set because of the rule
that allows only one strongest preference mark per ballot.

This method generalizes Smith Approval.

Forest