[EM] Why the concept of "sincere" votes in Range is flawed.

Abd ul-Rahman Lomax abd at lomaxdesign.com
Wed Dec 3 09:13:57 PST 2008

At 06:25 AM 12/2/2008, Kristofer Munsterhjelm wrote:

>That's not really what an approval cutoff is. An approval cutoff is 
>used by some methods to denote "the candidates above are those I can 
>accept; those below, I really don't like". At least that's what I 
>understand, though some methods may reward strategic placement of 
>the cutoff as well.

They must, actually. I.e., the voter may have some absolute and 
sincere thought, "I won't accept anyone but my favorite." But if the 
voter knows that sticking with this will make their vote moot, they 
must compromise somehow. And, by definition, this compromise is "strategic."

I haven't seen the idea mentioned before my recent posts, but a 
question that a voter might ask themselves is "Which would I prefer, 
to elect this candidate or to see a runoff take place under the 
rules? If the voter would prefer to elect the candidate, the voter's 
shift in approval cutoff that might result is a sincere answer to the question.

"Strategic" could be translated as "With consideration of the overall 
desires and welfare of the entire electorate, together with my 
personal preferences." In other words, "strategic" isn't bad. It 
distorts votes, however, and we'd prefer to encourage maximally 
accurate voting, because a good voting method can then truly optimize 
the result, *but* prohibiting or trying to prevent strategic voting 
could have the opposite of the desired effect, i.e., it could make 
results worse, not better. Remember, DSV can improve Range results. 
And so can honest voting. Indeed, I don't know the details of the DSV 
simulations, but if I recall my reading of the paper correctly, 
honest voters had the option of directing the automatic strategic 
voter to not optimize their vote. I don't know what mix of voters 
doing or not doing this was used, or how full-DSV -- every voter 
optimized -- compares to honest Range. (Again, "honest" here means 
accurate utility expression, which is, after all, not necessarily 
easy! It would mean that any significant preference is expressed, and 
that preference strengths were also expressed, which is the hard 
part. I've assumed that they are not expressed with full accuracy, 
that many Range Voters, intending to vote sincerely, will essentially 
vote a Borda ballot. equal preference strengths, or something close, 
unless there are very few candidates. They would also shift their 
votes to the positive half or negative half, which will likewise make 
preference strengths more accurate.

But I don't expect voters to wire themselves up to a polygraph and 
measure their emotional responses to candidates. We do not need fully 
accurate votes, as long as they are approximately right, because the 
votes will be averaged, effectively, over many voters.

(I'm truly warming up to positive/negative Range, with the default 
vote being midrange, i.e, 0, with max being +N and min being -N. 
Conceptually, it is much easier to understand. Range 2 is a piece of 
cake. (And, in that simple a method, 0 means "accept but not prefer 
and not strongly reject" ... In higher Range, the Accept level can be 
set off from zero.)

>The odd thing is that no party has actually done that. If using 
>Approval would let a party pick a candidate that's more likely to be 
>accepted by thevoters, then why wouldn't they? They could only gain. 
>The same holds if what the parties are really interested in is 
>finding candidates at +/-0.5, not +/-0.001 - a better election 
>method would let them do so more consistently, which would give a 
>better result (to them).

It's an interesting argument, with some merit, that it's a bad idea 
because nobody is doing it. However, Approval has been used for a 
long time. The Venetian Republic survived for 700 years, and used it 
as part of their election process for the Doge.

There is another explanation: existing leaders got their positions 
under the existing system. Make the system better, it destabilizes 
the power structure, there will be some changes. People don't like 
change that may take some power away from them, on average. (And, 
legitimately, they can be afraid that a newly empowered membership 
may make decisions unwisely, being inexperienced.)

The big obstacle to decent election reform is not, however, those who 
cogently oppose it. It is inertia and lack of interest. Until things 
get really bad, most people don't even think of change. They have 
other things to worry about. Many, many special interests benefit, at 
least they think they do, from the status quo. (That's not surprising 
and is not a condemnation of them; indeed, they are *adapted* to the 
status quo, they evolved to be able to use it and to benefit from it.)

But a rising tide raises all boats. In fact, better democratic 
methods will avoid such nasty phenomena as violent revolutions, class 
warfare, confiscatory tax structures, etc. They will *incorporate* 
the best of what exists, and not drastically remove power from those 
who have some legitimacy in exercising it. They will *use* the 
experience of the people who have been in control, they won't pull a 
Cultural Revolution and beat them, humiliate them, and send them to 
dig weeds. It does not harm society if some people have more wealth 
than others, it can benefit it, if the wealth is justly obtained and 
used. But neither should any special interest be allowed to harm the 
society as a whole in order to pursue its special interest.

Special interests are an essential part of the system, and will 
remain so. What is needed is only balance.

More information about the Election-Methods mailing list