Talk:IRV Prime: Difference between revisions
Content added Content deleted
mNo edit summary |
No edit summary |
||
Line 28: | Line 28: | ||
--[[User:Marcosb|Marcosb]] ([[User talk:Marcosb|talk]]) 16:41, 4 August 2021 (UTC) |
--[[User:Marcosb|Marcosb]] ([[User talk:Marcosb|talk]]) 16:41, 4 August 2021 (UTC) |
||
Was looking a little deeper at the theorem: |
|||
there must be a profile P arbitrarily close to this (in the proportions of ballots of each type) that does not yield a tie |
|||
The problem is, such a profile P may make it impossible for c to become the Condorcet winner; looking at all the profiles P where a wins (we must increment by 2 otherwise we continue to have a tie): |
|||
P1: |
|||
{{ballots| |
|||
abc: 5 |
|||
acb: 2 |
|||
bca: 3 |
|||
bac: 2 |
|||
cab: 3 |
|||
cba: 2}} |
|||
P2: |
|||
{{ballots| |
|||
abc: 3 |
|||
acb: 4 |
|||
bca: 3 |
|||
bac: 2 |
|||
cab: 3 |
|||
cba: 2}} |
|||
P3: |
|||
{{ballots| |
|||
abc: 4 |
|||
acb: 3 |
|||
bca: 3 |
|||
bac: 2 |
|||
cab: 3 |
|||
cba: 2}} |
|||
It becomes clear that in Profile P where a tie is broken & a wins, a wins because they are the Condorcet winner, so it cannot be true that c is the Condorcet winner. |
|||
--[[User:Marcosb|Marcosb]] ([[User talk:Marcosb|talk]]) 17:20, 4 August 2021 (UTC) |
|||
== Arrow/IIA == |
== Arrow/IIA == |