[EM] please read quick question about Condorcet

[Forest Simmons]

Another method that makes use of ballots showing relative preference
strengths, but this time a type of approval runoff.

For this method ideally there would be only one preference marker of each
strength.  If such is not the case, there are various ways of adapting the
method, including "symmetric completion" of the preference strengths,
analogous to the symmetric completion of ranked ballots in the case of
partial rankings.

At each stage of the runoff, on each ballot the strongest level marker
that is still straddled by at least one pair of the remaining candidates
is taken as the current approval cutoff for that ballot.

All candidates with lower than average approval are eliminated, and the
process is repeated until only one candidate remains.

Forest