New Entropia's Electionator 1.x

by **New Entropia** » Sun Jun 20, 2010 2:00 pm

I here introduce to you (and I am well aware of the irony of my nation being a dictatorship yet launching a spreadsheet to simulate elections) Electionator 1.0. Electionator is a spreadsheet program developed in Microsoft Excel to simulate the results of elections held under First Past The Post and Mixed Member Proportional Representation voting systems, as well as the Hare Quota method of Party List Proportional Representation.

Electionator will allow you to calculate the seats won by up to 7 parties (I know, its an odd number to pick. I'll increase it soon.) by entering the percentage vote share and the percentage of votes for each party that dont count toward the seats. That is, the percentage of the votes that only delivered a 2nd or 3rd placed result for instance, and are thus essentially wasted under First Past The Post. From this it will calculate the number of seats won, and will almost always deliver a majority government, as is the case with First Past The Post elections in real life. The spreadsheet will also allow for landslide results, such as the 1997 UK Election.

To demonstrate how Electionator works, I have set it up to simulate the UK's 2005 General Election. As you will see, if it is set up correctly then it produce pretty realistic results, its only a few seats short of the real result. I plan to add more electoral systems in the future (though probably not AV or STV) such as Supplementary Member system and Single Non-Transferable Vote.

To Download Electionator, click here:
Download (Excel .xls format, 219kb)

by **New Entropia** » Mon Jun 21, 2010 11:23 am

I should have written this up in the last post, but I kind of forgot. OK, here are a few notes on how to use Electionator.

Vote Share
The total vote share secured by each party. Because you need to enter the vote share you can still put up election threads in the NationStates forum and then use Electionator to calculate the number of seats.

Seats
The yellow highlighted fields in the seats column are the seats the party held before the election. Thus the number of seats before can be different to the number of seats up for grabs at the election. For instance, the 2001 UK election had 659 seats up for grabs, the 2005 election produced a 646 seat parliament. You must enter the number of seats the parties hold in these fields or Electionator will produce a weird result, or no result at all.

Wasted Votes
This is the percentage of votes that don't count toward seats, the votes that only achieve a second, third or fourth placed result. This can be a more difficult one to judge, and I will try to come up with some kind of formula to work this out automatically. However, as a general rule of thumb, no party will usually have less than 60% of it's votes wasted. An opposition that is on the up will usually have a lower percentage of votes wasted than the incumbent government whilst minor parties will usually have up to 95% of their votes wasted. Really, it's up to you. Have a play around with the percentages until you find a result that suits you and seems to reflect what you would percieve the result to be.

Landslides
You may have noticed that the results for the two biggest parties seem to fluctuate when a landslide result is produced. That is because the spreadsheet calculates landslide results (any result where the largest vote share is greater than 8% above the second largest vote share) employing a multiplying/dividing factor based on random numbers - hence the fluctuation.

Mixed Member Proportional (Also known as the Additional Member System, AMS)
This is basically a system where there are a certain number of MPs elected on a First Past The Post basis and a certain numer who are given seats on a party list basis depending on the vote share in the constituency seats. In the event that a party wins more seats than it is entitled to according to it's vote share, a situation with overhang seats occurs. Electionator assumes your MMP system assigns all parties seats in proportion to overhangs. If your system doesn't award overhangs then you just need to add together constituency seats and list seats.

If you don't know what MMP is, check the Wikipedia link in the post above.

by **New Entropia** » Fri Jun 25, 2010 1:16 pm

Update: Electionator 1.1 is out. In this version, you can select whether you are entering poll ratings or the percentage share of the vote the party wins. That is, you can enter in an opinion poll rating for that party and by use of random numbers, electionator will calculate the vote share that party wins. This allows for more randomness in elections and goes some way to simulating voter behaviour and showing how opinon polls are simply not an accurate reflection of how people actually vote. You can still enter a value for the vote share that a particular party wins if you so wish.

Download (Excel, 82 kb)

by **Orlkjestad** » Fri Jun 25, 2010 6:16 pm

Here's a nice idea:

-What about entering the percentage of people in the electorate that belong to a certain demographic, entering the likeliness of that demographic to vote for a certain party, and then calculating how many votes it got?
-Perhaps you should enter a new section for prime ministerial/presidential/gubernatorial races, since those don't involve seats, only electoral colleges?

by **New Entropia** » Sat Jun 26, 2010 8:44 am

That isnt a bad idea actually. I'll have a look at that. Elections for single seats as in gubernatorial/by elections etc would likely require more data to be input, the same going for presidential elections using an electoral college, but the ethnic group idea is a good one.

by **New Entropia** » Wed Jun 30, 2010 1:55 pm

Update: 1.2 is out. This version has multiple new features:

Added the ability to specify the likelihood of a particular demographic group to vote for a particular party
Added the ability to specify how your MMP system deals with overhang seats
Added the ability to simulate yes/no referendums based on the way the voter is leaning
Added more graphs

For the demographic group votes, I have set it up as default to use the A/B/C/D system of classifying sections of society used by UK government offices (again using 2005 election data). A/B are people who are employed in managerial roles, or executive positions, C1 are skilled manual workers, C2 are unskilled manual and D/E are long term unemployed, jobless, etc. This only works if you specify that you have entered the opinion poll percentage for each party and not vote share.

I am working on single seat elections such as presidency, by-elections etc, and I also plan to cater for more PR systems such as the Droop and Imperali quotas and the variable seat method.

Download (Excel.xls format, 138 kb)

by **New Entropia** » Sun Jul 04, 2010 12:20 pm

Update: 1.2.1 is out. This version is a minor update that fixes a bug in the formula for calculating FPTP elections. Other than the bug fix, 1.2.1 is the same as 1.2. The next main update, 1.3 will be out in a short while (I'm working on it now), and will definately include single seat elections and the number of parties will be increased to 10. For now however, download 1.2.1 here:

Download (Excel .xls format, 138kb)

by **New Entropia** » Thu Jul 08, 2010 1:20 pm

Update: 1.3 is out. This update incorporates the following changes:

Added single seat elections
Increased number of parties to 10
Increased maximum number of demographic groups to 10
Added ability to set vote threshold in Hare Quota PR
Added graphs, updated existing graphs to incorporate 10 parties

Download (Excel .xls format, 219kb)

by **Diol** » Sat Aug 21, 2010 2:11 pm

Your link is brokened!

by **Soyuz 11** » Sat Aug 21, 2010 3:54 pm

Diol wrote:Your link is broken!

Fix'd.

New Entropia's Electionator 1.x

New Entropia's Electionator 1.x

Who is online