On Wed, Sep 16, 2009 at 12:01 AM, James Gilmour <[email protected]> wrote: > The first statement seems logical, but I don't know about the second > statement. > I don't understand how an elected candidate could > be eliminated - sounds like a contradiction of terms.
Well, if you just re-run the count and exclude the withdrawn candidate, there is no guarantee that a candidate would be re-elected. For example, assume a 2 seater with votes 30: A>B>C>D 20: B>A>C>D 15: C>B>D>A 20: C>D>A>B 15: D>C>A>B Total 100 Quota 34 Result is Round 1 A: 30 B: 20 C: 35 D: 15 C elected Round 2 (1 vote transferred) A: 30 B: 20 C: 34* D: 16 D eliminated Round 3 A: 46 B: 20 C: 34 D: 0 A elected If C hadn't stood then the results would have been Round 1 A: 30 B: 35 C: 0 D: 35 B and D win. Thus A lost his seat because C withdrew. Ofc, with something like CPO-STV, this would be less of an issue. > I don't know what any of this means as I am not sufficiently familiar with > the inner workings of Meek STV. In Meek, elected means that you have at least a Droop quota (and can have any "keep value" between zero and one) . The method is designed so that no matter what changes happen to the "keep values", all elected candidates are guaranteed to have greater than or equal to the quota after the change. Each step has been checked and there is a proof that none of the steps in the procedure can change a candidate who has more than a quota into a candidate who has less than a quota. However, the convergence procedure will tend to bring the elected candidates closer and closer to the quota, but slightly above. However, increasing the keep value of eliminated candidates would take votes from the other candidates and potentially reduce a candidate who had more than a quota to a candidate that has less than a quota. ---- Election-Methods mailing list - see http://electorama.com/em for list info
