As much as I like the difference that was done here for city ruling council, from the other Ironrealm games, I still think it's missing.

How can we really conclude elections where 7 people or more contest, each get few votes and someone wins by having 1-2 votes more then the second.

What I'd like to see is elections being concluded only with a majority, lets say 51% votes goes to one person in a first round election and if not, get those in first and second place and re-contest them automaticly so one will definitly will have a majority over the other.

More the once I found it hard to decide between two people and think that this way gives a chance to find the best person people really want.

Interesting. I think that would be a very good idea. It would make it so more people support the winner. I second this

That is a great idea, sometimes its almost like everyone is running so there are very few voters!

This way if the large first runoff dosent get 51%, there would be a second runoff etc until the actual people's choice does win.

Right now anyone who spent a few days ratting can run, and that is ok but it doesnt make them qualified.

The problem with this is that it would take forever to elect somebody with the current election length. And if the election length was drasticly reduced, not everyone would have a chance to decide properly.

Perhaps there should be one more election after the initial one between the top two candidates? I know that in Serenwilde Nikua won by a few votes over one other, both of whom were far, far ahead of the other slew of candidates.

The problem I see here is peopel joing into elections just to draw votes and screw up the percentages.

I know people can probably already join to draw votes indiviually, but it would screw up percentages so much more.

I don't like the election system. It is at least as bad as the US election system in real life. In some ways, much worse.

What if people had to get X number of people supporting their challenge before could enter a contest? It would make people think before contesting and eliminate people who can't get together political ties and factions early on. Ideally, it would not necessarily force you to VOTE for them, but if at least, say, 10 people (for a city election) had to lend you their support before you could contest, you'd have fewer candidates, and those left would be the ones who had supporters in the first place...

the only problem I see with the election system is that a corrupt Guild leader can stay in office forever.

I honestly think this is a great idea. I have had many encounters prior that a city official has been elected even though they had a lower majority of votes. The votes were just so spread out that it was truly unfair as to what the citizens wanted in that position. I think percentage of votes and overall votes should both be taken note of. Popular vote should always win in these types of elections.

Sounds like what is need is ranked voting. You vote for all candidates in order of preference. This in RL avoids certain vote spliting which is based on having more than one candidate with some alike aspects or qualifications.

Now to do this by hand would be a pain, but when you vote they don't make you tally the votes of everyone else, so voting this way can be handled by computer. It is harder to explain then to do. ;-)

Not only is this a more accurate representation of who the voters want but it removes the need for a runoff election and permits small groups or candidates to have a visible presence. Most importantly it nullifies vote spliting or vote cancelling.

Consider the following election from a sampling of 100 votes.

Candidates
40 40% 1. Conservative in liberal city.
30 30% 2. Well liked liberal female DA.
15 15% 3. Well liked moderate.
10 10% 4. Liberal with a voting block (racial/gay/age/etc.)
5 5% 5. Liberal anarchist.

The result would indicate that the conservative candidate was a favorite.
~~~~

Now with ranked voting you vote your order of preference (I think Bob is the best choice, but if the rest of you don't like Bob, then I think Tom is better than Joe.) this is determined buy voting in order who you would like to see have the position.

1-10 represents a vote each. 5 the person they would most like, 1 the person they would least like.

1 2 3 4 5 6 7 8 9 10 (10% of 100 votes).
---------------------------------

4 5 3 5 4 3 5 3 3 4 = 39 ...1. Well liked liberal female DA.
2 4 5 2 3 5 4 2 4 3 = 34 ...2. Well liked moderate.
3 3 2 4 2 4 3 4 2 5 = 32 ...3. Liberal with a voting block (racial/gay/age/etc.)
5 2 1 3 5 2 1 5 5 2 = 31 ...4. Disliked conservative in liberal city.
1 1 4 1 1 1 2 1 1 1 = 14 ...5. Liberal anarchist.
Totals are not percents but point totals.

Now looking at the numbers after the = symbol it seems more the case that the concensus is that of the whole group the female DA is the best candidate. In addition, you wind up with a sense of how much suport in general any given candidate has.

Now this doesn't have to be that confusing when you vote. Something like the following isn't that hard for someone to use and there are dozens of ways the interface can be set up.

Bob, Tom, and Joe are running for office.

>Election xxxx vote
Who would be your first choice for X?
1. Bob
2. Tom
3. Joe
>election xxx vote 1
You cast your vote for Bob.

Bob is your first choice. Of the remaining, who would you like to see in office?
1. Tom
2. Joe

>election xxx vote 1
You cast your vote for Tom, leaving Joe as the candidate you feel is least qualified.

Or somesuch. As I said earlier, the math is for computers and need not be anything anyone voting needs to deal with.

Sounds like preferential voting, the system Australia uses in feredal elections... only in Australia it doesn't directly go to preferences if one of the candidates has enough votes.

Basically everyone votes for the candidates in order of preference. Say the candidates are Snickers, Twix, Bounty and museli...

Snickers -
Twix ------
Bounty ---
museli ----

So Snickers is your first preference, but if you can't have that you want Bounty, then museli, then Twix. All the votes are counted, and if one candidate doesn't outright have more than 50% of the vote (ie. 50% of the vote, then one more) they scrap whichever candidate got the least votes, look at the second preferences for whoever voted for that last-placed candidate and assign them, then keep doing that until one of the candidates has 50% +1 vote. Say things look like this after the first counting:

Snickers: 10 votes.
Twix: 35 votes.
Bounty: 40 votes.
museli: 25 votes.

Under the most-votes-wins (aka. 'first past the post', or 'direct election' system, I think it's called) Bounty would win. But they need 50% of the votes, +1, so they get rid of Snickers as a choice because they have the least votes, look at all the ballots of people who voted Snickers as number one, check their secondary preferences, and assign them. So after that it might look more like:

Twix: 42 votes.
Bounty: 41 votes.
museli: 27 votes.

Still none of them have the clear 51 votes, so you go again, this time looking at the museli preferences and assigning those:

Twix: 62 votes.
Bounty: 48 votes.

Though Twix originally wasn't winning, all the preferences mean that people get a bit of a say even if their #1 hope isn't elected. Um. It can seem like a complicated system and I probably didn't explain it too well, have a look at section 9.1 of this .pdf file from the Australian Electoral Commission website.

Preferential voting is different to the standard 'whoever gets the most votes' system, but as Australia has more political parties (despite being a predominantly 2-party system) than America's basic republican vs. democrat system it's more appropriate. Ballot cards basically always have candidates from at least three, if not four or five, different parties.

(Then there's the Senate system which is a whole other matter and is WAY more complicated... twelve Senators per state, and two per territory, only they elect half of the senators from the states each time at three year intervals... and that has no relation to anything that could be applicable for a nice sensible game like Lusternia so we won't get into that.)

Um, so considering city elections have more than two - typically four or five, as I think you need to have at least three people running for an election to start at all - candidates, a preferential system might be more appropriate. It means that people who vote for the candidate who comes last haven't totally thrown their votes away, and also gives the winner a clearer mandate to rule (imagine an election where the winner only got 35% of the votes, and the other two people running got 34% and 31% of the vote each, it's not a very clear victory and leads to people later down the track complaining).

Just a random small fact that turned into a not-so-small post... if you think preferential voting would be a better idea than the current majority wins system, feel free to discuss it.

edit: Um, so some other people already suggested this, just not using the term "preferential voting", above... sorry for being so unobservant! But I hope I explained it a little more clearly at least.

The ideal solution is to implement the Condorcet system. Plurality ("first past the post" that the US uses) voting is indeed one of the most mathematically flawed systems, second only to instant run-off (or even the normal run-off proposals above). The problem with Condorcet is that even though it really is not that complicated, it requires some thinking to understand and will confuse players who will assume it's some complex mathematical formula. Perhaps approval voting could be done, which is the next best thing, according to the math.

There is a great site that analyzes the game theory mathematical properties of different voting systems, including Borda (frequently used in football), but I lost it. But for example one problem with run-off systems is that it is mathematically possible to cause someone to lose by voting for him or to cause someone to win by withdrawing a vote for him. Plus it has all the same problems as plurality voting with splitting the vote and "defensive voting" for the lesser evil. It is not as obvious, though. The moment the minor party becomes an actual threat to the two main parties, the defensive voting syndrome sets in. So it only gives the illusion of ending this problem so long as the minor party has no chance of winning.

Of course no election system will be mathematically perfect. The Arrow Impossibility Theorem proves this.

Isn't it possible to end up with a broken cycle, under Condorcet? As in, A is voted better than B, B better than C, and C better than A?

As for sites, I know there's some decent information on voting systems (and most everything else) in wikipedia.

That is correct. There are two ways to address it.

1) After eliminating all the candidates that are defeatable, use an inferior voting method to break the cycle , like IRV or Borda. Those do not require a re-vote since they use the same input.
2) That situation is not as bad as it seems. It's merely the Condorcet version of a tie. So treat the election as a draw and do whatever it does right now with a draw.

This is a pretty good document on different voting systems:

http://bcn.boulder.co.us/government/approv...te/altvote.html

I've referenced it a couple times.

Use a preferential voting system. It would not be too hard to run if you only had a dozen or so candidates running.