[EM] More about Summed-Ranks (SR)

[Michael Ossipoff]

I'd like to add something. Ordinary Borda has a particularly
undesirable kind of clone problem, like that of Copeland. Factions
have incentive to nominate many candidates. Even when the number of
alternatives is fixed, sets of very similar alternatives are favored.


In Summed-Ranks (SR), that problem is alleviated by the
power-truncation. In typical Borda, a faction who have nominated lots
of candidates have an advantage, but in SR, their large number of
candidates is penalized on ballots that bottom-rank them.


And, as I said, SR is Approval for any voter who uses only the ranks
of top and unranked.


SR can be justified only if it has an easy handcount, only if its
handcount is much easier than that of ICT and Symmetrical ICT. Such is
the case.


SR's count-labor is proportional only to the first power of the number
of candidates. It's only necessary to make one pass through each
ballot.


Of course pandering to laziness, by providing rank balloting, isn't in
every way the most desirable goal. But, in practice, in polling, it
can be tempting to make voting as easy as possible for voters, to give
the simple 1sst choice, 2nd choice instruction. And, in discussions of
voting systems for official public elections, some people prefere rank
balloting.


SR achieves rank balloting with a much easier handcount than the
Condorcet methods, including ICT and Symmetrical ICT.


Of course Approval and Score are preferable to SR, in addition to
having easier and simpler counts.


In a poll or public election, SR's method-chosen ratings for the
various rank postiions can be justified as approximations to how the
voter might have rated in score, while letting the voter not have to
rate the candidates hirself. So it could be regarded as an
approximation to Score's SU maximization under sincere voting, or to
Score's fractional ratings due to uncertainty or for SFR.


Mike Ossipoff