[EM] Even simplier anti-gerrymandering rule

[raphfrk at netscape.net]

 Initially, the nation should be divided into districts such that the number of districts is not less than (10?) times the number of seats to be allocated and are contiguous.
 
 These districts cannot be modified after the initial allocation, except once when new census information is collected according to these rules
 
 - the threshold shall be the one tenth of the total population per seat to be allocated
 
 - a district must be divided into 2 contiguous districts if it has a population greater than double the threshold, such that the 2 resulting districts have populations larger than the threshold. The boundary that divides the district into two parts shall not be longer than one and a half times the shortest possible boundary that would divide the district into two valid parts. If no valid boundary is provided by the legislature, the SC shall use the shortest boundary to divide the district. This rule shall also apply to the resulting districts.
 
 - two or more districts may be combined into a single district as long as the resulting district has a population less than the threshold. The resulting district must be contiguous.
 
 - no other changes are permitted
 
 (an alternative rule to the above would work too. Basically, these districts should be hard to change. However, they should also be kept alot smaller than the final districts).
 
 The above districts shall be known as the Initial Districts
 
 The election districts shall be decided as follows:
 
 1) The districts shall be set to the Initial Districts
 
 2) If the number of districts is greater than the number of seats to be filled, combine the two neighbouring districts that have the smallest total population into a single district and goto 2)
 
 
 This works even better for PR-STV or asset voting type districts
 
 1) Set the districts equal to the counties
 
 2) Distribute seats among the districts using the d'hondt method
 
 3) If any district gets fewer than (4) seats, combine the two neighbouring districts which have the lowest total population into one district and goto 2)
 
 By increasing the number of seats in step 3, greater voter to seat ratio accuracy can be achieved.
 
 Also, the smaller the populations of the Initial Districts, the closer the resulting districts would be to well balanced.
 
 Another option would be to use boundary length as part of the algorithm.
 
 Like:
 
 2) If the number of districts is greater than the number of seats to be filled, combine the two neighbouring districts where the resulting district has the smallest total boundary length and where the resulting district has a population less than the largest district. If no such combination is possible, combine the two neighbouring districts that have the smallest total population into a single district and goto 2)
   
 I think there may be issues near the end with the scheme though as the districts will be large. For example, if you had 10 districts at around 10% of the population and 9 seats, then combining any 2 would give one district at 20%.
 
 Maybe there would need to be a second pass to balance out the sizes. This is less of a problem is the districts are multi seat districts.
 
 Another option is to use the multi-seat distribution method with a smallest size district of 1 and apply the algorithm recursively if any of the districts end up with more than 1 seat (and keep going until all districts gets at least 1 seat).
 
 
  Raphfrk
 --------------------
 Interesting site
 "what if anyone could modify the laws"
 
 www.wikocracy.com
________________________________________________________________________
Check Out the new free AIM(R) Mail -- 2 GB of storage and industry-leading spam and email virus protection.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20061106/5363b19b/attachment-0002.htm>