[EM] SL vs LR--Rounding is unavoidalbe because allocations are integer

[Juho]

Yes, also time could function as the balancing tool. If we talk about  
seats allocated to parties then changes in the program of the party  
could cause some (minor) confusion. A party that has changed its  
policy radically since the last election would get seats based on the  
votes that it got when it had another policy.

Another form of time division would be to allocate different  
candidates of a party terms of different length. Let's say that a  
party that has 100 candidates gets 10 seats. It divides the seats to  
30 most popular candidates (this is an open list election). Each seat  
(with fixed term length) is divided between three candidates in time  
so that everyone's time in office corresponds as closely as possible  
to the number of votes she got. All candidates will get a positive or  
negative value depending on if they got more time than they got votes  
(probably 30 candidates will get a negative value and 70 candidates a  
positive value). Next election will take these values into account.  
The algorithm will be something similar to what you proposed.

Juho Laatu


On Dec 10, 2006, at 23:32 , raphfrk at netscape.net wrote:

> > From: juho4880 at yahoo.co.uk
> > On Dec 10, 2006, at 20:50 , MIKE OSSIPOFF wrote:
> > > But
> > > rounding is quite unavoidable, since fractional seats can't be
> > > given (or at least are against the rules).
> >
> > I agree. All methods lead to rounding errors (unless we cut the
> > candidates in fractions or give them unequal voting power :-).
>
> What about charging each State 1 seat for every seat they are
> allocated.  If a State receives to few seats in one election,
> they will receive a compensating one in a future election.
>
> The long term average number of seats allocated to the State
> would be exactly proportional.
>
> For example:
>
> Each State has a seat total that is not cleared from election to
> election.  This total counts in fractional seats.
>
> When a seat is allocated, the State that receives the seat's total
> is decreased by 1 seat.  All States, including the one that
> received the seat, then have their total increased by State
> Population divided by National Population.
>
> A seat is always allocated to the State with the highest total, or
> using some tie-break rule if there is a tie.
>
> The sum of all the State's totals will always equal zero.  A
> State with a positive score has been under allocated seats based
> on its population, and the State with the highest score will thus
> be most entitled to the next seat.  I don't think it is possible
> for any State to exceed +1, but am not sure.
>
> One issue is that the totals are likely to get a bit hairy as they
> are fractions.  It might be worth specifically setting an accuracy
> required (say 1/1000 of a seat).
>
> Another issue is that there is a certain amount of randomness.
>
> Finally, handling small States would require a kludge.  Perhaps,
> make a rule that they must be allocated a seat at the end, but that
> it isn't included as part of the totals.  This would mean that
> sometimes they would get a seat directly and sometimes they would
> get a seat due to the exception to the normal rules.
>
> A different option would be to allow States to form super States.
>
> This would be like the idea to allow parties to form sub-groups.
>
> Two States might be entitled to 4.4 and 5.8 seats each.  This would
> give them 4 and 5 seats each and 1 seat that is common to the 2  
> States.
> Probably when voting the votes for the 2nd State would count for  
> more as
> the seat is more closely theirs.
>
>
>
> 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.
> ----
> election-methods mailing list - see http://electorama.com/em for  
> list info

Send instant messages to your online friends http://uk.messenger.yahoo.com