[EM] SODA might be the method we've been looking for.

Jameson Quinn jameson.quinn at gmail.com
Thu Dec 15 08:05:38 PST 2011


2011/12/15 Andrew Myers <andru at cs.cornell.edu>

> On 7/22/64 2:59 PM, Andy Jennings wrote:
>
>> I don't see any huge theoretical downsides.  Do others still have
>> reservations about SODA?  I realize that some people may be opposed to
>> delegation, in principle.  And others think delegable systems just don't
>> have a chance of getting implemented.  So I think these debates about which
>> is the best voting system in the standard (non-delegable) model are still
>> useful.  I also think it's useful for Jameson to inject a plug about SODA
>> every now and then.
>>
> I will repeat what I've written before:
>
> > I have to agree. SODA to me seems quite complex. It appears to pose
> difficult strategic decisions for candidates and even for voters.
>
> Thanks for the honest response.

What do you think would help alleviate this largely-false appearance? Voter
strategy is limited to a few cases:

1. Correct approval strategy in case your favorite candidate's preferences
differ significantly from yours. People on this list understand approval
strategy; in my opinion, it's not ideal, but it's no worse than plurality
strategy, which most people tolerate. And I estimate that perhaps a third
or fewer voters will differ significantly from their favorite candidates.
If "significantly" only counts differences in the order between the two
frontrunner candidates, that kind of number makes sense.

2. Attempts at "chicken" strategy in a few cases. In the classic A+B vs C
case, such strategy can only work if C has no preference between A and B.
(Under one rule variant of SODA, even an honest preference that wasn't
predeclared would be sufficient to avoid a chicken dilemma). Note that,
unlike in approval/Range/MJ, the only way a chicken strategy can work for A
is by making it impossible for B to win the election; chicken strategy is *
always* either ineffective or dangerous. So it seems to me that in SODA,
unlike those systems, there is no slippery slope to a chicken dilemma.

As for candidate strategy, that comes in two flavors:

1. Preference declaration strategies. Again, these mainly come down to
chicken strategies, and there are several restraints even on such
strategies. If A truncates B, B can retaliate; this should keep it from
happening unless A is clearly a second-string candidate, in which case it
may be a good thing. Also, C could intervene to avoid the dilemma.

2. Post-election strategy. This is a sequential, perfect-information game;
there's a single optimal strategy, and in any real election it's pretty
easy to calculate. (I can imagine artificially-balanced situations with
dozens or hundreds of candidates which might be NP-hard; but in real life,
it basically comes down to finding the delegated CW).

Also note that journalists would quickly work out and publish the optimal
strategy and all plausible variations thereof, so the candidates would not
have to work it out on their own.

So, I can't quite give a blanket denial that strategy matters, but I can
give a qualified one: in real life SODA elections, it is not worth worrying
about strategy. Having read the above, can you see any way I could say that
better? I want to be able to allay this concern; strategy issues are an
outstanding strength of SODA, not a weakness.

Jameson
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20111215/aa438fbc/attachment-0004.htm>


More information about the Election-Methods mailing list