More standards

[David Marsay]

In response to:
> To:            Election Methods 
> From:          Blake Cretney <bcretney at my-dejanews.com>
> Subject:       More standards

> Here are the standards I nominate.  They are not in order of my
> preference.
> 
> Party manipulations
I agree with these.
 
> 3.  Marginal majority

I agree with the sentiment, but I am not clear about the precise 
criterion.

> Definition of Sincere Voting

> 
> 4.  Sincere vote matches tabulation
> The best way of judging this I can think of is to ask, "is a sincere
> vote a rational vote assuming no knowledge of how others are voting."

This is too complex. To be rational, a voter must be trying to meet 
some objective. We suppose that voters have a preference order and 
try to get the best result they can.

The obvious criterion is that they rank candidates according to their 
true preference. We need something like this.
Variants:
They rank their most preferred candidate first.
They rank honestly, but may truncate
They rank honestly, but may tie (e.g., Approval).

> 5.  No weighting
> It would be better if a voters sincere vote was not only justifiable,
> but the best in every way.  That means, for example, that a voter
> should not have to decide between increasing the chance of getting
> their first choice elected, and increasing the chance of getting
> their first or second choice elected.  Except for strategy based
> on how others are voting, the sincere vote should be best in all ways.

What does this mean, apart from no strategic truncating or tieing? 
LO2E is more specific and hence more useful as a criterion. 

> Justifying the method as a search for best candidate

> 5.  Non-reversible
> If we assume that our method is finding the best candidate, and all
> ballots are reversed, so that A > B > C becomes C > B > A, then we
> are finding the worst candidate.  If our method can sometimes
> find the same candidate for best and worst, we have reason to
> believe it isn't really finding the best candidate.

What about votes:
10000 ABC
100  BAC
1000  BCA
1000  CAB
1000  ACB
10000 CBA

I think a method that gives B is quite reasonable. A failing of FPP 
is that it can do this unreasonably. Hence I believe that we should 
never elect a candidate whom a majority rank last.
 
> 6.  Monotonic
Agreed, but challenging
--------------------------------------------------
Sorry, but apparently I have to do this. :-(
The views expressed above are entirely those of the writer
and do not represent the views, policy or understanding of
any other person or official body.