[EM] AV strategy is hard and that's good

[RLSuter at aol.com]

Some of the criticisms expressed about Approval Voting (AV) don't strike me 
as very reasonable. One person objected to AV in a private message because he 
believes it suffers from the Prisoner's Dilemma problem. But the leading 
academic advocate fo AV and one of its co-inventors, Steven Brams, is an expert on 
game theory and surely understands the Prisoner's Dilemma as well as anyone, so 
that doesn't seem a reasonable objection.

Other people have said that AV favors strategic voters. But none of the 
arguments to that effect have seemed persuasive to me, because they have failed to 
take into adequate account the many different preference orderings and 
strengths that different voters have and the fact that information about the 
preference orderings and strengths of other voters is so difficult to know and to 
account for. It seems to me that for many voters, even the most strategically 
sophisticated ones, it is often impossible to know what the best strategies are 
for maximizing the impacts of their ballots according to their particular 
preference orderings and strengths. I wonder if it is not the fact that AV makes 
strategic voting so difficult to figure out that it works so well and does so 
well at overcoming the Prisoner's Dilemma problem for voters as a group.

To help illustrate the strategic difficulties AV can pose for even 
strategically sophisicated voters, imagine that in the 2004 election, Bush and Kerry had 
both been polling about 35% prior to the election and Nader had been polling 
25%, with 5% undecided or for other candidates. In this scenario, it is very 
conceivable that Nader would win an AV election if many Kerry voters regarded 
him almost as highly as Kerry and much higher than Bush and a smaller 
percentage of Bush voters regarded him almost as highly as Bush and much higher than 
Kerry. Call the first group Kerry-Nader voters and the second group Bush-Nader 
voters. If the pre-election polls are assumed to be very accurate, both 
Kerry-Nader voters and Bush-Nader voters would have to be concerned about the 
possibility of a very close election with either Bush or Kerry coming out ahead. 
Given that, I suspect a significant number of both Kerry-Nader voters and 
Bush-Nader voters would vote for Nader as a hedge to help ensure that if their first 
choice doesn't win, their second choice (Nader) will. But it's a difficult 
choice for them to make. There's no way for any voter, no matter how strategically 
sophisticated, to know what the best way to vote is. Their hedge vote could 
prevent their first choice from winning (a bad outcome) but it could also 
prevent their last choice from winning (a good outcome).

AV is not my favorite method, and I'm not advocating it here for presidential 
elections. But I don't think it's nearly as bad a method as many people have 
been saying, and it does have one distinct advantage. It's very easy to tally 
an AV election. No new voting equipment is needed. It can be done easily with 
even hand counted paper ballots. It can also be done easily and quickly in 
meetings where voters must choose from among three or more options. This happens 
a lot, as for example when an organization is trying to decide where the next 
meeting should be or a new organization is trying to pick a permanent name. 
Even for presidential elections, given the many problems with current election 
equipment, an AV vote with hand counted paper ballots would have been a vast 
improvement over how the 2004 election was actually carried out.

-Ralph Suter