[EM] 05/21/02 - FWD Math for those who want more Math:

[Donald Davison]

05/21/02 - FWD Math for those who want more Math:

Below is more election math for Gosh and Poor Richard, both of whom have
asked for more math on this list.

Enjoy, Donald,

  ------------ Original Math Letter ------------
X-From_: research at ijs.co.nz  Sun May 19 20:23:29 2002
X-Sender: research at 202.89.128.27
Date: Mon, 20 May 2002 12:23:30 +1200
To: Mike Ossipoff <nkklrp at hotmail.com>
From: Craig Carey <research at ijs.co.nz>
Subject: Archery of Mr Ossipoff: missing IRV always
Cc: Donald Davison <donald at mich.com>
Mime-Version: 1.0


 >
 >That "burying" problems of where this happens:
 >
 > [1]  added paper = 1(BC...), and B wins
 ><-->
 > [2]  added paper = 1(ACB...), and C wins
 >
...

Change that to:

To Mr Ossipoff:

-----------------
A method passes the test if this never occurs:

  [1]  added paper = 1(B...),  and B wins and A loses
<-->
  [2]  added paper = 1(AB...), and A and B both lose

(No constraints on the number of winners and the number of candidates).
Using my notation, this is being rejected:
    "Prohibit: (B+)<-->(A-B-)"

-----------------

Instead of talking about burying in general, you'd be less ignored by
EM members if avoiding the generality of burying.
Note to Mr Davison: if there is even a tiny failing there then it is
something voters would want to get around sometimes, e.g. if alert to
the problem. The recent advice at EM from Mr Davison did not concede
that. It is whether the essential manipulation is of preceding
preferences or of trailing preferences (like Approval requires). The
problem could partly go away if everyone does their fullest to
manipulate public elections and there was a decent method in use.

 >
 >Craig Carey


I am not subscribed to that EM list.


----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em