[There's an irony here which is probably worth noting. In past discussions, Anthony Towns had been in favor of eliminating all options which don't beat the default option early on in vote resolution, rather than including any special treatment of the default option in any iterative part of the vote resolution process. And, I had been opposed.]
Background: http://electionmethods.org/Arrow.html Arrow defined a set of criteria which an ideal voting system should have, and then proved that no voting system could satisfy all these criteria. A part of his model is that only the winner of the vote is relevant. The tough votes to decide are those where cyclic ambiguity exists. [For the current proposed ammendment: those where the schwartz set contains more than one member.] There's two ways of resolving cyclic ambiguities: [1] Use some mechanism to pick from among the ambiguous choices [2} Discuss further and vote again When the majority of voters agree that something should be done, it's pretty clear that [1] is better than [2]. When the majority of voters do not agree that something should be done it's pretty clear that [2] is better than [1]. The default option is our mechanism for distinguishing between these two cases. Note that the default option cannot be cloned: options which rank below the default option are "bad options", options which rank above the default option are "good options". Thus, we SHOULD ignore voting criteria which expect to treat the default option as if it were any other option. Thus, to some degree, any of the options which beat the default option are acceptable options while none of the options which are beaten by the default option are acceptable options. Finally, here are the vote tallies for our past non-leader votes. In essence, this shows that the default option has not been a popular option, and (more generally) that the proposed changes to the voting system would not have affected the outcome of any of these votes. To read these tables: rows represent votes for an option, columns represent votes against an option. Thus, if you see a 5 if the second row and the third column that means that there were 5 ballots which explicitly preferred the second option to the third option. constitution 0 86 86 "y - Yes" / winner 0 0 5 "n - No" 0 58 0 "f - Further discussion" logo1 0 37 77 "s - SINGLE License" 69 0 85 "d - DUEL Licnese" / winner 20 17 0 "f - Further Discussion" logo2 0 15 37 22 18 24 22 39 "a - ANTS" 103 0 104 89 92 77 84 104 "w - SWIRL" / winner 35 16 0 22 20 17 24 38 "e - SEAL" 88 47 85 0 69 54 64 86 "o - OLD" 72 36 73 46 0 43 48 78 "c - FIXED CHICKEN" 87 57 91 69 74 0 65 90 "d - DG" 67 36 67 58 54 54 0 72 "m - MODIFIED" 56 27 61 32 33 32 38 0 "f - FURTHER Discusson" logo3 0 89 96 "s - FOR logo swap" / winner 20 0 38 "a - AGAINST logo swap" 9 44 0 "d - FURTHER Discussion" [The constitutional vote is the only one that would reasonably have a supermajority requirement. Here, the winner was unanimously preferred over all alternatives.] Looking at the logo votes: vote winner beat winner beat status quo by next most preferred option by logo1 4.25:1 1.86:1 logo2 3.85:1 1.35:1 logo3 10.7:1 4.45:1 Here, all of the non-default options were classified as significantly better options than the default option. FYI, -- Raul
