[EM] Yee Diagrams for Bucklin Voting

[Leon Smith]

Ka-Ping Yee's code has weaknesses,  but it is straightforward.
Sadly,  it only supports 4 candidates.  I implemented 4-candidate
Bucklin ranked 1 through 4.    Honestly I'd rewrite the whole thing,
but this was a quick and easy way to see how Bucklin behaves.   I
included the four scenarios that Yee has on his page in the zip file I
sent,  although with 20,000 voters and not 200,000.

Out of curiosity,  has anybody contemplated algorithmic improvements
for generating Yee diagrams more quickly?   For example,  moving the
distribution a small delta would typically change the votes of a small
number of voters,   so it should be worthwhile to intelligently update
votes and totals.   Although I imagine this kind of improvement would
benefit some methods  (such as Condorcet and Bucklin) moreso than
others,  (such as IRV).

Certainly,  the algorithms are going to have to get better if you want
a reasonable interactive 2-d experience.

Best,
Leon

On Tue, Jan 25, 2011 at 6:29 AM, Raph Frank <raphfrk at gmail.com> wrote:
> How are you handling the ranks?
>
> Is it 4 ranks?
>
> one first ranked
> one second ranked
> as many 3rd ranked
> rest 4th ranked
>
> Deciding the 3rd rank requires similar strategy as approval and could
> change how the diagrams look.
>