[EM] voting methods

[Dave Ketchum]

On Jun 5, 2009, at 12:58 AM, Árpád Magosányi wrote:

> I guess the list might have opininons in this discussion.
>
> 2009/6/4 Warren Smith <warren.wds at gmail.com>
> > I am using "winning strategy" in the game theory sense.
>
> --voting is then an N-player game with N very large.
> Most or all attempts to look at voting in that way have been  
> unsuccessful.
> Game theory is best suited to 2-player games and most of it breaks
> down with more than 2 players.
>
> I am not an expert in game theory, but I feel that the existence of  
> term "tactical voting" shows that it can say things about elections.  
> I always hear political analysts talking about best strategies for  
> parties in an election campaign. Also there are a lot of voting  
> criteria concerned with what happens when candidates do things (e.g.  
> bring in clones), or voters do things (modify ballot in some or  
> other way, or do not vote).
>
> > Range voting also have a problem with tactical voting. As tactical  
> voting
> > is
> > the winning strategy for voters, it boils down fast to approval  
> voting,
> > which is not much better than plurality voting.
>
> --Approval voting seems to be quite comparatively good if all the
> voters are tactical.
>
> In particular, if all voters are tactical then plurality, IRV, and
> Condorcet systems all
> are EFFECTIVELY THE SAME (i.e. all elect the same winner) in  
> scenarios where
> there are 2 major-party candidates plus also some minor party
> candidates whom the voters believe have little chance to win.  That
> is, more precisely, if the voters
> always rank one major top and the other bottom (tactically) then a
> major will always win in these voting systems, and a minor can never
> win.  Over time this effect causes the
> minor parties to die and we get 2-party domination.
>
> And game theory enters here. With some Condorcet dishonest voting  
> gives unnoticeable advantage to the voter. With range vote it does  
> much, and eventhe  ballot leaves more room for it. Plurality even  
> punishes honest voting. So voters with nearly anything but Condorcet  
> will reduce the entropy of the ballot by tactical voting, and lead  
> the system to two party system. Because the winning strategy in a  
> two party system strengthens it.

This needs more careful thought:

In Plurality with majors expected to win, only those voting for a  
major get counted in deciding on winner - others might as well not vote.

With Condorcet, and even Range, voters can apply full effort to the  
duel between majors whether they do or do not expect such to win.   At  
the same time, they can vote for third parties, with this effort  
measurable and having potential to win if strong enough.

IRV, with luck and the same voting, can do like Condorcet; with bad  
luck can horrify.

Dave Ketchum
>
> > From an information science perspective it is clear that a  
> preferential
> > ballot have more information than an approval ballot.
>
> --and range ballots still more.
>
> Not, if you count the effect of tactical voting: range ballot  
> effectively becomes an approval ballot.
>
>
> > Why do we discuss it in private anyway?
>
> --OK with me if you post this email.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20090605/74a39e6b/attachment-0004.htm>