If anyone wants to see what the revised stats look like in graphical form, see below. I've converted percentages into votes by multiplying by 453 (the number of voters) and ordered by the number of first choice votes.
Ok, so kramerc has gone through the statistics for the rounds stats and preference percentages, he's found a few errors with the calculations, which should now be fixed in the latest statistics.
The quota is actually 56, not 57 and the number of valid votes was 443. The 453 figure was how many people actually voted, but 10 of those had invalid votes. None of this has affected the results of the winners, just the statistics showing how they won.
Wait.. how cam tom numbers be this off compared to last time.. what happen to SkylordLuke's votes? Kepticks votes changed little? How can their be this large of differences with these two stats. Seriously look fishy. I don't understand how it could be messed up. It should be simple, I don't know how you go from 3.53 to 14.67.....in first votes. And when this first results came out, they were justified as he won with the second, third and etc. Wtf
[DOUBLEPOST=1430311673,1430311411][/DOUBLEPOST]
Ok, so kramerc has gone through the statistics for the rounds stats and preference percentages, he's found a few errors with the calculations, which should now be fixed in the latest statistics.
The quota is actually 56, not 57 and the number of valid votes was 443. The 453 figure was how many people actually voted, but 10 of those had invalid votes. None of this has affected the results of the winners, just the statistics showing how they won.
well i suppose since tom had enough to get in in just the first votes alone, as well as skylord, and nearly dalmont and comrade as well, you had his votes roll down the hill and finally go out the window entirely because everyone who voted for him also voted for another winner. 4 of the seats were taken up in the first round of voting alone lol
I think there may still be a bug in the round-by-round stats. The total in each row should only ever decrease from 443 as some people lose their 5th choice and their vote gets discounted entirely, but the totals are increasing sometimes as well.
Would have to agree with Draconis, something is still off with the round-by-round. The obvious way of seeing the error would be in the final round, in which 6 people have 56 votes and 1 has 63. Under normal STV no one should be over the set requirement to get elected (56, which is ~1/7th of all people voting), and because some votes end up being discounted, there should be at least 1 or more persons who isn't quite at the required 56.
Mayhaps it was a rounding error? When votes get transferred, I recall hearing they were distributed as %s, and with 44 rounds you can amass quite an inaccuracy.
Would have to agree with Draconis, something is still off with the round-by-round. The obvious way of seeing the error would be in the final round, in which 6 people have 56 votes and 1 has 63. Under normal STV no one should be over the set requirement to get elected (56, which is ~1/7th of all people voting), and because some votes end up being discounted, there should be at least 1 or more persons who isn't quite at the required 56.
Mayhaps it was a rounding error? When votes get transferred, I recall hearing they were distributed as %s, and with 44 rounds you can amass quite an inaccuracy.
There's 1 person with 63 because he was the last guy that got selected, there is no need to continue the rounds for those excess votes since all 7 council members are selected.
There's 1 person with 63 because he was the last guy that got selected, there is no need to continue the rounds for those excess votes since all 7 council members are selected.
I went back over it, apparently I missed the part were the votes were reduced before being divided by 7. When I was reading about STV, it said the threshold was 1/7 of all votes (so 56*7= ~ total #votes), so I never noticed that adjustment.
Looks like there was indeed a bug with the round stats, the increasing in total votes was caused by votes being over-transferred (transferred more than once) to their 4th and 5th preferences in the statistics. This has now been fixed.
We've also added totals to the latest statistics for each round, so you can see what is happening with elimination. I've also decided to include another statistic that shows dead votes, instead of removing dead votes, they are just listed and go nowhere.
Here are the latest (and hopefully correct this time) stats:
kramerc has checked the algorithm as well, just in case. A few things were improved and some minor bugs fixed. These stats represent the latest recalculation of winners and the latest stats calculation.
Yeah, this 6 page discussion on votes could have easily been avoided by popular vote. If someone doesn't get votes then they should campaign better. Popular vote = 1 vote per person. There is no screwing it up with algorithms or fancy stat charts. 450+ votes one with the most wins. Simplest form of voting.
Yeah, this 6 page discussion on votes could have easily been avoided by popular vote. If someone doesn't get votes then they should campaign better. Popular vote = 1 vote per person. There is no screwing it up with algorithms or fancy stat charts. 450+ votes one with the most wins. Simplest form of voting.
Yeah, this 6 page discussion on votes could have easily been avoided by popular vote. If someone doesn't get votes then they should campaign better. Popular vote = 1 vote per person. There is no screwing it up with algorithms or fancy stat charts. 450+ votes one with the most wins. Simplest form of voting.
Except campaigning is very unfair when a few people here have youtube channels that can instantly get their message across to their thousands of subscribers. Who cares if 6 pages of this topic have been posted? It's been proven that it works, as was shown in the latest stats. As long as it works we should be satisfied.
Except campaigning is very unfair when a few people here have youtube channels that can instantly get their message across to their thousands of subscribers. Who cares if 6 pages of this topic have been posted? It's been proven that it works, as was shown in the latest stats. As long as it works we should be satisfied.
Agreed, many people were afraid it would be nothing more than a popularity contest. It may be more complicated but if it does produce a more fair result it should be worth it. Also this is the very first council so we have a bugs and procedure to work out.
I've made a little gif animation showing the round by round stats and how the votes get redistributed as people are eliminated if anyone is interested.
The automatic election threshold is the dotted black line, and the 'dead votes' bin collects votes that cannot be re-distributed. Click on the thumbnail to enlarge.
I've made a little gif animation showing the round by round stats and how the votes get redistributed as people are eliminated if anyone is interested.
The automatic election threshold is the dotted black line, and the 'dead votes' bin collects votes that cannot be re-distributed. Click on the thumbnail to enlarge.
I've made a little gif animation showing the round by round stats and how the votes get redistributed as people are eliminated if anyone is interested.
The automatic election threshold is the dotted black line, and the 'dead votes' bin collects votes that cannot be re-distributed. Click on the thumbnail to enlarge.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
