[EM] Is strategic voting a bad thing, really?

[wclark at xoom.org]

> Are there cases where the winner according to strategically cast ballots
> is a better choice than the winner according to sincere ballots?

Here's one such example, assuming the CW is a better choice, in cases
where IRV elects someone else...

Sincere preferences:

A>B>C:25
A>C>B:14
B>C>A:15
B>A>C:14
C>A>B:10
C>B>A:22

Condorcet: B wins

A>B:25+14+10=49
B>A:15+14+22=51
A>C:25+14+14=53
C>A:15+10+22=47
B>C:25+15+14=54
C>B:14+10+22=46

Sincere IRV: A wins

A:39,C:32,B:29
A:53,C:47

Strategic IRV: B wins

A>B>C (same)
A>C>B (same)
B>C>A (same)
B>A>C (same)
C>A>B -> A>C>B
C>B>A -> B>C>A

A>B>C:25
A>C>B:14+10=24
B>C>A:15+22=37
B>A>C:14

B:51,A:49

Yes, there are different strategies that could be tried, such as A>C>B
changing to C>A>B ... but then A>B>C would simply switch to B>A>C and B
wouls still win.  Perhaps the fact that B is the CW guarantees that there
will always be a strategy that allows B to win under IRV as well?  That
would be an interesting fact, if true.

Also, this is obviously an unrealistic scenario in that organizing such
large-scale strategic voting is infeasible in the real world, due to
limited information.  HOWEVER -- if a full theory of optimal strategies
could be worked out, it might be interesting to devise a two-stage
election method that utilizes sincere ballots as inputs, then determines
what the "correct" strategic ballot would be for each voter (assuming
perfect information.)

-Bill Clark

-- 
Ralph Nader for US President in 2004
http://votenader.org/