[EM] Optimal Cardinal Proportional Representation

[Kristofer Munsterhjelm]

On 2024-05-05 17:22, Richard Lung wrote:
> As I understand it (or misunderstand it) cumulative voting  is an early 
> version of cardinal voting going by quantities of points or "plumping" 
> for a single candidate. Of course the details are different but they 
> share the similarity of cardinal voting obscuring an ordinal vote, a 
> straight-forward order of choice, in the hands of the voters, instead of 
> the parties. ultimately reducing to FPTP under the plumping option.
> 
> Or as HG Wells would say: enfeebling complications fruitful of corruption.

That does seem pretty accurate, but I don't think the observation can be 
generalized to all cardinal methods.

The problem with cumulative voting is that you have a given number of 
points and need to spread them around. If you don't plump them all for a 
single candidate, you're making your vote weaker. Hence it behaves like 
FPTP.

But with approval or Range, you can set your rating of each candidate 
independently, so how much of your voting power you can give to A 
doesn't depend on how much you've given to B. So the incentive to plump 
isn't nearly as strong.

There are definitely weaknesses to cardinal voting. But cardinal methods 
aren't all the same: different rules give different weaknesses.

-km