[EM] Margins vs. Winning Votes

[Dave Ketchum]

On Sun, 24 Jul 2005 23:27:07 +0300 Juho Laatu wrote:

> Hello Dave,
> 
> On Jul 22, 2005, at 17:25, Dave Ketchum wrote:
> 
>> On Wed, 20 Jul 2005 21:36:00 +0300 Juho Laatu wrote:
> 
> 
>>> - In raking based real life elections it seems to be quite common 
>>> that voters don't give full rankings. This example has only three 
>>> candidates and therefore full rankings could be quite common. But the 
>>> election could have also considerably more than three candidates, in 
>>> which case partial rankings probably would be quite common. It is 
>>> probable that ranking candidates of competing party is less common 
>>> than ranking candidates of ones own party (just like in this example).
>>
>>
>>
>> Partial rankings ARE appropriate - you properly do not rank when you 
>> do not know or care which of the leftovers are better.
> 
> 
> I agree. I just try to study when partial rankings are used. In the 
> example candidates of the competing party were unknown to some of the 
> voters, which led to leaving them out of the ballots. The example had 
> only three candidates (to keep it simple). In real elections number of 
> candidates may be also much higher. In this case many voters certainly 
> would find ranking all of them (especially the unknown ones) tedious. 
> Partial rankings are thus very natural, and leaving unknown and 
> uninteresting candidates and unlikely winners unranked are probably some 
> of the most typical cases.


For all of this, partial rankings should NOT be penalized.  Pressuring 
voters to rank beyond their interest or understanding makes voting more 
painful, and can result in ranking that does not represent what a voter's 
desires might have been if the voter had done more thorough preparation.

> 
>>> Now, what if some of  the the 20 C supporters (C>B voters) would note 
>>> the weak position of C before the election and decide to vote 
>>> strategically C>A>B.
>>
>>
>>
>> The word "strategically" turns me off.  For whatever reason, they are 
>> switching between C>B>A (with A's position implied) and C>A>B - 
>> between B is preferable to A to A is preferable to B.
>>      Not surprising that B no longer wins.
>>      Not surprising that C wins in WV - C retains its ranking while B 
>> becomes less desirable.
>>      Perhaps less surprising that A wins in margins - these voters 
>> were SUCCESSFUL in controlling outcome between A and B (though they 
>> were unable to cause C to win).
> 
> 
> I made the assumption (although I didn't mention this in my mail) that 
> the original votes in the example were sincere, so the changed votes 
> were not in line with the voters' true preferences. I don't know in what 
> sense word "strategically" turns you off, but for me the problem is the 
> unwanted property of a voting system where some voters may make their 
> favourite win by strategical voting. And what makes the "strategic 
> problems" of this example worse is that it seems that as a result of the 
> strategy the winner is a candidate that voters didn't seem to value very 
> much (different opinions on the strength of C may however exist).


"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.

> 
> Technically the findings are not surprising but at lest for me the fact 
> that this insincere voting technique can be applied in some pretty 
> normal election situations with reasonable probability of success and 
> low risk (and when a "not so good" candidate wins as a result) sounds 
> like an unwanted feature that should be studied. Maybe the existence of 
> this problem is not a big surprise since it is known that Condorcet 
> methods are not strategy free but there is maybe some element of 
> surprise in that it is the winning votes that seem to suffer from this 
> problem. It seems to be a general belief that winning votes are less 
> vulnerable to strategies than margins. I presented this example to shed 
> light to the winning votes vs. margins discussion also from the opposite 
> angle. I'm eager to hear if the experts can dig out some conclusions out 
> of this example or estimate its importance.


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.

> 
> Yours,
> Juho

-- 
  davek at clarityconnect.com    people.clarityconnect.com/webpages3/davek
  Dave Ketchum   108 Halstead Ave, Owego, NY  13827-1708   607-687-5026
            Do to no one what you would not want done to you.
                  If you want peace, work for justice.