It is true that there are numerous possible metrics for rating voting systems and that makes it impossible to order the voting systems linearly (in a line from worst to best) that will agree with all reasonable metrics. But the same can be said about the problem that voters face.
Voters apply a variety of metrics in comparing different candidates and often this must be the case because voters have limited knowledge about the candidates and often what they know about one candidate is something they do not know about the other, so that metric has to be avoided in making the comparison. An just as fit is impossible for voting experts to establish a linear order on voting systems, for the same reason it is impossible for voters to establish a linear order on candidates when there are many candidates. That makes gives a big advantage to cardinal methods like score voting a big inherent advantage when it comes to both accuracy and simplicity.
But there is not much difference, by nearly any metric, between voting systems for elections with only two candidates. That is why a voting system that does little to encourage there being more than two candidates is a poor alternative. Breaking up the duopoly of two big parties should be the primary qualification for consideration of a better voting system. Ranked-choice voting fails to meet this objective and so does approval voting. What is needed is a system like balanced approval voting which is evaluative like score systems are but which are balanced to allow a vote against a candidate as easily and as effectively as a vote for the candidate (here I mean by a vote, just one of the many preferences specified on a ballot; it should not be interpreted as the entire complex of preferences indicated on the ballot). Balanced approval voting would seem to be the simplest possible voting system that is both evaluative and balanced.
It is true that there are numerous possible metrics for rating voting systems and that makes it impossible to order the voting systems linearly (in a line from worst to best) that will agree with all reasonable metrics. But the same can be said about the problem that voters face.
Voters apply a variety of metrics in comparing different candidates and often this must be the case because voters have limited knowledge about the candidates and often what they know about one candidate is something they do not know about the other, so that metric has to be avoided in making the comparison. An just as fit is impossible for voting experts to establish a linear order on voting systems, for the same reason it is impossible for voters to establish a linear order on candidates when there are many candidates. That makes gives a big advantage to cardinal methods like score voting a big inherent advantage when it comes to both accuracy and simplicity.
But there is not much difference, by nearly any metric, between voting systems for elections with only two candidates. That is why a voting system that does little to encourage there being more than two candidates is a poor alternative. Breaking up the duopoly of two big parties should be the primary qualification for consideration of a better voting system. Ranked-choice voting fails to meet this objective and so does approval voting. What is needed is a system like balanced approval voting which is evaluative like score systems are but which are balanced to allow a vote against a candidate as easily and as effectively as a vote for the candidate (here I mean by a vote, just one of the many preferences specified on a ballot; it should not be interpreted as the entire complex of preferences indicated on the ballot). Balanced approval voting would seem to be the simplest possible voting system that is both evaluative and balanced.