[EM] Margins vs. Winning Votes

[Juho Laatu]

On Jul 25, 2005, at 01:24, Dave Ketchum wrote:

> "Strategically" still turns me off.  Voters who preferred B over A, 
> and had planned to vote accordingly, are gambling that they can get 
> better results by claiming, instead, to prefer A over B:
>      In some cases they can, unfortunately, succeed at what they claim 
> to want.
>      If these can change their votes, then so can others for other 
> reasons, thus destroying the knowledge all used in plotting.
>      How cycles operate was used in the plotting.  Cycles are valuable 
> in resolving near ties, but deliberately setting them up for intended 
> results is tricky.

My thoughts around the pair-wise comparison based voting methods and 
strategies are roughly as follows. Rating based methods are maybe too 
vulnerable to strategies in contentious elections, so we may have to 
satisfy with ranking based methods. Pair-wise comparison methods are a 
wonderful solution that allows voters to express their preferences 
quite extensively (but still easily) and in most cases without any 
strategy problems. Sincere voting thus seems to pay off (better than in 
many currently used election methods). Unfortunately we have the cycle 
related problems (otherwise the methods would be quite flawless). 
Natural cycles are luckily quite rare. Artificially generated cycles 
are a bigger problem. The big question is if pair-wise methods are 
still good enough to be recommended for use. I think the differences 
between different tie breaking methods are not that big. Since all of 
the pair-wise comparison methods have some of the basic 
vulnerabilities, the more important question seems to be if the 
pair-wise comparison methods are useful in general as a group. In real 
election situation also the real life environment has an influence. If 
majority of the voters vote sincerely, the strategic voting problem 
remains just as background noise in the process without causing any 
considerable harm (((hmm, the example I presented studied the the 
possibility that very few strategic votes could pick a "bad 
winner"...))). If the method and attitudes of voters are bad enough, 
then strategic voting might be a problem. And if the impact of 
strategic voting would be worse than what currently used election 
methods have (e.g. some really bad candidates would be elected), then 
we would need to deem pair-wise election methods unusable (is some 
environments?) and would be forced to go back to (recommending) some 
simpler election methods (e.g. IRV, two round runoff, plurality, 
approval). So far I'm leaning in the direction that strategic voting 
would at least in large public elections be marginal (or done mostly by 
voters who didn't know that in the new method they don't need to vote 
strategically but the a sincere vote is likely to defend their 
interests in the best possible way). I'm also leaning in the direction 
that voting methods that pick the best winner could be more useful than 
ones that aim at eliminating strategies. Partly because people want the 
method to provide best possible results, partly because the strategic 
fixes may be marginal, partly to keep the method simple and 
understandable, and partly because strategic countermeasures easily 
lead people to thinking that strategic voting is a key property of the 
method. I'm not 100% sure that the simplest methods work fine and I'm 
still waiting for someone to prove that some methods are unusable and 
some usable (but haven't seen (or understood :-) that yet). So, for me 
the question is if the basic pair-wise comparison methods are strategy 
free enough to be used as practical election methods as they are. And 
my guess is, yes they are in most single winner elections.

> But here the example you constructed resulted in a win by the plotters 
> in wv - and in the result they claimed to want, but which was worse 
> for them than what they would have achieved without plotting, under 
> margins.

Yes, based on this single isolated example one should recommend margins 
based methods and not winning votes based methods. I in general do like 
margins also since they seem to be a more natural measure of preference 
than winning votes, so this result is just fine for me :-). Winning 
votes have maybe some burden of being overloaded with hopes of finding 
a panacea in the fight against strategic voting.

Best Regards,
Juho