Talk:Cube rule

Law of large numbers?
I don't see how the Cube Rule is related to the law of large numbers. Could someone explain? Thanks. --Allen 01:36, 20 March 2006 (UTC)
 * that was the Kendall Stuart argument. See Rjensen 01:42, 20 March 2006 (UTC)
 * I don't believe so. AFAIR, it has more to do with the fact that in modern industrial democracies the geographical distibution of blue-collar versus white-collar workers has a variance of about 13%. see Seats, Votes and the Spatial Organisation of Elections, Gudgin & Taylor (1979); and Geography of Elections, Johnston & Taylor (1976). RodCrosby 02:37, 20 March 2006 (UTC)


 * The law of large numbers reference doesn't appear to be meaningful and should probably be removed. The law of large numbers states that the average of a sequence of random samples will converge to their mean provide "enough" samples are taken. It does not involve a formula and is only a statement about long term behavior, hence there is nothing to really result in a cubic approximation. A reference to the central limit theorem would be slightly more appropriate because here at least the normal curve is involved. In, it is stated that Kendall and Stuart considered voting constituencies (i.e. districts) with normally distributed support for a particular party. Apparently, if a certain standard deviation is used, then as support for one party changes, there will be a cubic relationship between a parties average level of support and the number of constituencies that are above/below the 50% threshold. Other ways in which the Cube Rule can be derived are also discussed. EERac (talk) 15:02, 6 November 2008 (UTC)

Three-party politics
The article appears to be self-contradictory with respect to the validity of the rule for three party systems. Brondahl (talk) 12:44, 8 May 2010 (UTC)

The 2002 election
I don't see how the cube rule was relevant in 2002 election. The vote ratio was 1.105:1, but the ratio of seats is only 1.117:1 instead of 1.348:1. 195.148.159.170 (talk) 18:06, 15 January 2016 (UTC)