From kragen@dnaco.net Fri Aug 28 15:39:52 1998 Date: Fri, 28 Aug 1998 15:39:51 -0400 (EDT) From: Kragen To: clug-user@clug.org Subject: voting procedures Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Keywords: X-UID: 1560 Status: O X-Status: There are several voting procedures that produce better results than simple majority vote when deciding on more than two candidates. If we were to vote on the new CLUG name with simple majority vote, it's possible that most people (say 60%) would prefer almost any name other than CLUG, but their first preferences might differ widely -- CGLUG, CFSUG, CUUG, etc. Even if 60% of the voters preferred "CGLUG" to "CLUG", we might end up with 40% of the vote going to "CLUG" and 20% each going to "CGLUG", "CFSUG", or "CUUG". So we'd pick CLUG in a simple majority vote. There is no perfect voting procedure, but here are a few that are better than simple majority vote. (I implemented the Borda count and Condorcet voting in Perl this morning, in hopes that one or more of them might be used tomorrow, and can trivially implement approval voting.) - The Borda count. Each voter ranks their preferences: "First CGLUG, then CFSUG, then CUUG, then CLUG". We assign 0 points to their last choice, 1 point to their next-to-last choice, 2 points to their third-from-last choice, etc, up to N points for their Nth-from-last (i.e. first) choice. The points from all the voters are totaled; the choice with the most points wins. In a slight variation, which I guess is more common, you can not bother to write your last few choices, and no points go to them. This has the advantage that you don't have to think about all of the choices, just your first few. The Borda count is subject to manipulation by "strategic voting" and agenda manipulation, but a lot of people think it's the fairest method, particularly with honest voters. - Approval voting. Each voter votes "yes" or "no" for each choice. The choice that gets the most "yes"es wins. This is very popular in many places. - Condorcet voting. Votes are rankings, as in the Borda count. Then each pair of candidates (A,B) is compared in a runoff; rankings in which A came before B are counted as votes for A, and rankings in which B came before A are counted as votes for B. If a candidate wins all their runoffs (with all other candidates), they have won. This method of voting has much to recommend it. It's not subject to "strategic voting" and agenda manipulation. It's impossible to "split the vote" or "waste your vote". Its biggest problem is that it doesn't always produce a result at all. If five people vote "CLUG, CGLUG, CUUG", five people vote "CGLUG, CUUG, CLUG", and five people vote "CUUG, CLUG, CGLUG", then each of the three wins one runoff and loses one runoff. - Single transferable vote. Each voter ranks their choices, as in the Borda count. The votes are assigned to their first choices, unless they go over some "quota" I don't quite understand; the candidate with the fewest votes is eliminated, and their votes are transferred to the next candidate down on the voters' preference lists, and the process is repeated until only one candidate is left (or as many as the number of seats that need to be filled.) With multiple seats, this tends to produce approximately proportional representation, and it's less susceptible to strategic voting and agenda manipulation, and vote "splitting" and "waste" is not as big a problem. Variants of this method are also known as "Hare voting", "Preference voting", "choice voting", "Fractional STV", and other names. Many governments use them. - Plurality with runoff. Most of the US uses this system. If no candidate gets a majority (>50%) in the initial election, the two candidates with the greatest minorities have a head-to-head election. This keeps a minority candidate from winning over a single strong majority candidate, but only ameliorates the plurality method's flaws a little bit. Well, that's plenty. I'd like to elect the Condorcet winner, if there is one, in our CLUG-renaming 'election'. Other thoughts? Kragen -- Kragen Sitaker We are forming cells within a global brain and we are excited that we might start to think collectively. What becomes of us still hangs crucially on how we think individually. -- Tim Berners-Lee, inventor of the Web