[EM] Scenarios defining the method space
Kevin Venzke
stepjak at yahoo.fr
Wed Feb 23 20:11:09 PST 2011
Hi,
Just a few notes. I tried to use the generator to explore the region
between DAC and DSC, but I didn't see much of value, so instead I looked
for certain scenarios which predictably separate the north from south,
etc.
Here's a map that shows the north-south-east-west orientation I want to
refer to. A method is located where its left-most letter is.
........................Bucklin.............................
............................................................
............................................................
..............................DAC...........................
...................WV.......................................
............................................................
............................BklnVar.........................
..............C//KH.........................................
............................................................
.............KH.............................................
.................C//IRV.....................................
.......................QR...................................
...................................DSC......................
..................IRV.......................................
.......................................FPP..................
The east-west dimension is narrower and fittingly I only see two
scenarios that define it:
1. A. BA. CB.
2. AB. BA. CB.
Eastern methods elect A on these, and western methods elect B. Bucklin
is split. Condorcet methods will all have to be on the west. Not sure
what to say about the east except that all the Participation-satisfying
methods are there.
For the north-south dimension:
3. A. BC. C.
4. AC. BC. C.
Northern methods elect C and southern methods elect A.
5. AB. BC. C.
Northern methods pick B and southern methods pick A.
The boundary is a little rougher with these three. The KHs and C//IRV
test ambiguously. The LNHarm methods are all in the south(*), and it seems
clear that minimal defense methods will all be in the north.
(*) except "MMPO,FPP" not pictured, which is more like a northern method.
There are other scenarios that give certain results only in a certain
corner.
My next project, I think, is to see whether I can get an automated
description of strategies for each method and voting faction. To some
extent I already have this (in pass/fail form) but I'd like numbers.
Kevin Venzke
More information about the Election-Methods
mailing list