Log in

No account? Create an account

Previous Entry | Next Entry

Slapdash post, from a comment I made elsewhere:

Eh, there's tradeoffs. I actually live under STV for Cambridge (MA USA) city election; having to rank 9-25 candidates can be a real cognitive pain, I find. (There's 9 seats, so you have to rank at least that many to have a full voice, and we had 25 candidates last election.) And 9 seats from a district means a threshold of at least 11% for a faction to get a distinct voice; if a group is spread evenly as 8% around the country, too bad.

Conversely, closed party list might as well be proxy voting for the party leaders, save for the dim possibility of revolt, and open part list gives you some control over specific candidates but you're still voting for a party group, not make-your-own-list in STV. Much simpler to vote for though, pick a candidate or party, bam you're done.

And thing is, STV extracts a lot more information from you but doesn't do much with it, since all that ranking precision just controls how your vote trickles through the count, without making things that much more representative in the end. To invoke math, ranking 25 candidates means 25! (factorial) or 1e25 possibilities, but the final result is simply choosing 9 out of 25, which is a much smaller number (2 million, or 2e6). Conversely, party list converts a simple vote directly into a party percentage.

And frequent referendums a la Swiss democracy would give me much more direct influence on the laws that pass, for still far less work than ranking 25 candidates, most of whom won't win... Voting in 9 referenda has me picking out of 2^9 options, and there are 2^9 possible outcomes of 9 laws passing or failing. Again, the work involved is directly proportionate to the result achieved.

See the comment count unavailable DW comments at http://mindstalk.dreamwidth.org/393508.html#comments


( 1 comment — Leave a comment )
Jun. 5th, 2014 01:44 pm (UTC)
I think you overestimate the work necessary in a coupe of ways.

Factorial is misleading because you can use your preference relation to build a binary search tree, providing a ranking with only n log n comparisons on average; 25 candidates is an easily doable ~116 comparisons.

Another bogus estimate is the exponential referenda; 2^9 work is the worst case of all 9 bearing materially on the same issues but being unconstrained enough to have no particular necessary correlation between answers. Most cases would just be nine binary choices.
( 1 comment — Leave a comment )


Damien Sullivan

Latest Month

November 2018


Page Summary

Powered by LiveJournal.com
Designed by Lilia Ahner