STV put simply -- Droop and Hare compared - effect of use of MMDs

A bare-bones explanation of STV is multi-member district where each voter has one vote. Votes are placed on candidates. Quota is determined and anyone attaining quota is elected. As well, if necessary, surplus votes of quota-winners are transferred according to next usable marked preference. If necessary, least-popular candidate are eliminated and their votes transferred. Process continues until all seats are filled by quota winners or until field of candidates is thinned to number of remaining open seats.

(or as you said regarding IRV, if necessary, "Repeatedly eliminate the candidate who [has] fewest [votes]."

that part of IRV is identical to STV. -

only other difference between IRV and STV is: STV is multi-winner which means surplus vote of winners need to be transferred, which happens only in relative small number of rounds of counting.

sure there are various ways to conduct transfers of surplus votes. just as there are various ways to construct open-list PR, which is often overlooked in "simple explanations" of list PR.

bearing in mind the DM, STV is just as proportional as list PR, if you use votes as placed at end.

with each winner having identical vote tally, or close to it for those elected in last round, there is no way this could not be true.

of course transfers can affect the parties' tallies of votes. that is part of why they are used. so using first-round votes to assess STV proportionality is not fair.

You say "Voting for that candidate counts as a vote for your party, but also a vote for that candidate, in the determination of which candidates get the seats won by your party. The candidates are seated in the order of their vote-counts. Any unfilled seats are then assigned to the remaining candidates in the list, in their listed-order."

but that actual allocation of seats to individuals is not part of your "simple explanation of St-L."

Droop is not biased in favor of large parties. it is just more in favor of them than Hare, which is tougher on them (because it allows surplus votes to rest with early winners when large parties want their votes shared out best to take more seats).

Droop is lowest figure that allows no more to get quota than the number of seats. anything more than that is un-necessarily large and may (or may not) affect the parties' seat counts.

where DM is large, difference between Hare and Droop is less than one percent of votes.

in DM-10, Droop is 9 percent plus 1, Hare is 10 percent.

so not much impact on large parties' rep.

at 21 DM as used in NSW Aus. the difference is even smaller-

Hare 5 percent

Droop 4.54 percent,

hardly going to make much difference when parties are thousands of votes apart.

big difference is scientific based system where candidates' vote tallies can be compared, and the most popular elected whether by taking quota or by relative plurality at the end.

use of MMDs means electorate broken not at all or into only a few large chunks.

and candidates within each chunk being compared to each other.

instead of one candidate getting a lucky break and being elected with only 1 percent of the votes overall while in another district winner has 5 percent of votes overall.

things like that happen under FPTP.

======================================================================

few PR systems above city level use at-large districting.

most list PR systems use DM of less than 20, which is perfectly workable under STV as well.

=====

Michael Gallagher Comparing P.R. Electoral Systems. Quotas, Thresholds, Paradoxes Majorities

common DM under PR:

7 is about average for DM used in Belgium, Denmark, Greece, Iceland, Norway, Spain and Switzerland

14 is about average for DM used in Finland, Italy, Luxembourg, Portugal and Sweden p. 485

=======

but any use of MMD decreases that shattering of the electorate, and decreases chance of unfair results.

illustration:

FPP 36 districts

PR (list or STV) perhaps 4 districts of nine, hopefully based on pre-existing geographic units --cities, counties, etc..

still some districting but much more fairness and more flexibility.

and when population shifts, simply add or take away one seat,

no re-districting required.

plus or minus ten or 20 percent from district to district much more flexible by thousands of votes when MMDs used, but also any variation is spread over more seats so appears more equal and fair.

say 1M voters 36 seats

FPP: median district size 28,000 so acceptable range is 22,222 to 33,600

MMD DM-9: median 252,000 so acceptable range is 202,000 to 302,000

=======================