Blaugust 31/31: 2015 Senate Elections

I was in two minds as to how to finish-up Blaugust, with a round-up style post or the long-awaited results on the tie election, but in the end the ties won the day. Let me just say how much fun it has been to be a part of the big blaugust thing again, and throw out a big thanks to everyone who dropped by becase of it. Now on with the show...

If you're not familiar with the back-story here, we've just run an election using the Australian single transferable vote system to appoint six ties from my somewhat eclectic collection to be the Leaflocker's reprsentatives in my suitcase as I go overseas, there being a limited amount of space given that I also need to take other less exciting things like underpants and shirts. Readers were asked to designate there preferences on which ties could represent them from a field of sixty-six, and today is the fateful day when we find out the results. If you're from a region that uses a less nuanced electoral system, sit back and enjoy the wonders of an STV system, the wonders of which I could natter on about until the cows come home if you gave me half a chance.

There were twenty-six ballots cast in the neck-and-neck election, and all were valid, so a big thumbs up to everyone for ticking all the boxes (or not, as appropriate). These ballots were spread relatively evenly between the original google form (11 ballots) and the later fancy voting application donated by the ever-brilliant Alecat (14 ballots) ,with one postal vote, and a majority of voters elected to cast their preferences for individual ties below-the-line (18) in preference to my quickly-assembles parties of ties above-the-line (8). Most notably, eight people filled in each and every position on the below-the-line ballot and three met only the minimum requirement of designating at least six preferences, whereas in the above-the-line ballot everyone designated full preferences, which interests me greatly and makes me keenly anticipate the more democratic results that should occur when we allow above-the-line preference voting in the near future (the Greens are tabling a bill here in South Australia next week, and undoubtedly the federal system will get there eventually once they get around to responding to the latest review of potential electoral reform).

Of the sixty-eight eligible ties, fourteen attracted at least one first-preference vote. Only the Christmas party failed to attract a vote, which I am going to take as a vote of confidence in the tie collection in general instead of taking it as criticism of the party system in general. Interestingly, not a single one of the second candidates on a party ticket got a look-in, but shout-outs go the Crimson Peacock, which despite not making the cut made the top five ties of almost every below-the-line voter, and undoubtedly would have been selected if we were using some kind of weighted voted system. The ties that got first-preferences are:

[CON 1] Polytetraflouroethtielene (5)

[DAD 1] Electric Boogaloo (3)

[DEM 1] The Travelator (3)

[LIB 1] Carribean Silk (2)

[LIB 3] Green Sakura (2)

[UNC 1] Australian Orienteer (2)

[CAR 1] Oh Tin Tin! My Tin Tin! (2)

[LIB 5] Crimson Tide (1)

[DAD 3] All Aboard (1)

[DOT 1] The Amicable Vicar (1)

[DOT 4] Fade to Blue (1)

[DOT 6] Prince of Persia (1)

[CAR 4] Mouses (1)

[PAI 1] Popular Penguins (1)

Since there were 26 valid ballots, the number of votes required to elect a tie to the suitcase (the quota) is 4.7 votes (Number of ballots divided by one more than the number of available spaces in the bag, plus one) [Note: Yes, I'm aware that we'd normally round the quota down to an integer, but we're dealing with a small number of ballots here, so I'm leaving it at one decimal point]. Since the leader of the conservative party, Polytetraflouroethtielene, attracted 5 first-preference votes, we already have a Senatie-elect. 

[CON 1] Polytetraflouroethtielene elected in position 1 with 5 ballots.

And of course, we have the joys of surplus-redistribution! Because every vote has to have the same value in an STV system, since more votes were collected than were required for the winning candidate, all of the votes are redistributed to their second preferences at a small percentage of their original value. This is where the maths gets fun.

0.18 votes (3 ballots at 0.06 transfer value) originally from [CON 1] Polytetraflouroethtielene distributed to [CON 2] The Gopher
0.06 votes (1 ballot at 0.06 transfer value) originally from [CON 1] Polytetraflouroethtielene distributed to [LIB 1] Carribean Silk
0.06 votes (1 ballot at 0.06 transfer value) originally from [CON 1] Polytetraflouroethtielene distributed to [LIB 4] Bad At Tetris

From here all the candidates that have attracted no preferences have been removed from the ballot. For ease of administration, I made up a ballot sheet with the sixteen top ties (the fourteen that attracted first-preferences and the two extra that attracted second-preference votes). If I was a clever programmer type I'd have written a program to do this, but there's something reassuring about having a big pile of ballot papers, isn't there? Now we begin that process of excluding those ties that have the fewest votes from the ballot and the redistributing the preferences. From this point on, preference numbers will describe preferences as distributed amongst this short-list of votes, not the total pool of candidates.

[DAD 1] Electric Boogaloo (3)

[DEM 1] The Travelator (3)

[LIB 1] Carribean Silk (2.06)

[LIB 3] Green Sakura (2)

[UNC 1] Australian Orienteer (2)

[CAR 1] Oh Tin Tin! My Tin Tin! (2)

[LIB 5] Crimson Tide (1)

[DAD 3] All Aboard (1)

[DOT 1] The Amicable Vicar (1)

[DOT 4] Fade to Blue (1)

[DOT 6] Prince of Persia (1)

[CAR 4] Mouses (1)

[PAI 1] Popular Penguins (1)

[CON 2] The Gopher (0.18)

[LIB 4] Bad At Tetris (0.06)

[LIB 4] Bad at Tetris excluded.
0.06 votes originally from [CON 1] Polytetraflouroethtielene distributed via 3rd preference to [LIB 5] Crimson Tide.

[CON 2] The Gopher excluded.
0.06 votes originally from [CON 1] Polytetraflouroethtielene distributed via 3rd preference to [LIB 5] The Travelator.
0.06 votes originally from [CON 1] Polytetraflouroethtielene distributed via 3rd preference to [LIB 1] Carribean Silk.
0.06 votes originally from [CON 1] Polytetraflouroethtielene distributed via 3rd preference to [DOT 1] The Amicable Vicar.

[DEM 1] The Travelator (3.06)

[DAD 1] Electric Boogaloo (3)

[LIB 1] Carribean Silk (2.12)

[LIB 3] Green Sakura (2)

[UNC 1] Australian Orienteer (2)

[CAR 1] Oh Tin Tin! My Tin Tin! (2)

[LIB 5] Crimson Tide (1.06)

[DOT 1] The Amicable Vicar (1.06)

[DAD 3] All Aboard (1)

[DOT 4] Fade to Blue (1)

[DOT 6] Prince of Persia (1)

[CAR 4] Mouses (1)

[PAI 1] Popular Penguins (1)

Now, someone unfamiliar with the system might wonder why we bothered distributing preferences to candidates that we then immediately excluded, but tracking the preference flow has had a couple of important effects. First of all it's given The Travelator the boost that it needed to become the outright leader, but more importantly it has shielded two of the bottom candidates, Crimson Tide and The Amicable Vicar, from potential exclusion until at least all the candidates that weren't so fortunate as to attract those surplus preferences.

In a normal Australian election, when the bottom candidates have the same number of votes who is ultimately eliminated relies on one being randomly selected from the worst-performing remaining candidates. However, since we're dealing with 26 votes instead of thousands, this is potentially a much more common event, and since it would be sad for random chance to influence the whole election at this early point (if at all), we require a system for selecting the least popular of the potential candidates for elimination. Thus I've implemented a count-back system for these tie situations (hurr hurr), relying on two additional numbers. First is the number of voters who included the candidate on their ballot, which should disadvantage ties which a larger percentage of the electorate chose not to nominate at all ,effectively voting no confidence in these candidates (it turns out that by some chance ordering exclusions by this number actually had no effect on the outcome of the election, but we did it anyway just in case). A second level of tiebreaker is the sum of the remaining preferences, but I don't feel like explaining that right now, so let's not even go there and hope that we don't need to.

Including the preference numbers so that we can progress further, the election looks like this:

[DEM 1] The Travelator (3.06) [21]

[DAD 1] Electric Boogaloo (3) [19]

[LIB 1] Carribean Silk (2.12) [20]

[LIB 3] Green Sakura (2) [18]

[UNC 1] Australian Orienteer (2) [20]

[CAR 1] Oh Tin Tin! My Tin Tin! (2) [21]

[LIB 5] Crimson Tide (1.06) [19]

[DOT 1] The Amicable Vicar (1.06) [18]

[DOT 6] Prince of Persia (1) [21]

[CAR 4] Mouses (1) [20]

[PAI 1] Popular Penguins (1) [19]

[DOT 4] Fade to Blue (1) [18]

[DAD 3] All Aboard (1) [17]

[DAD 3] All Aboard excluded.
1 vote originally from [DAD 3] All Aboard distributed via 2nd preference to [CAR 1] Oh Tin Tin! My Tin Tin!

[DOT 4] Fade to Blue excluded.
1 vote originally from [DOT 4] Fade to Blue distributed via 2nd preference to [DEM 1] The Travelator

[PAI 1] Popular Penguins excluded.
1 vote originally from [PAI 1] Popular Penguins distributed via 2nd preference to [DOT 1] The Amicable Vicar

[CAR 4] Mouses excluded.
1 vote originally from [CAR 4] Mouses distributed via 2nd preference to [UNC 1] Australian Orienteer

[DOT 6] Prince of Persia excluded.
1 vote originally from [DOT 6] Prince of Persia distributed via 4th preference to [CAR 1] Oh Tin Tin! My Tin Tin!

[LIB 5] Crimson Tide excluded.
1 vote originally from [LIB 5] Crimson Tide distributed via 4th preference to [CAR 1] Oh Tin Tin! My Tin Tin!
0.06 votes originally from [CON 1] Polytetraflouroethtielene distributed via 5th preference to [LIB 3] Green Sakura

[CAR 1] Oh Tin Tin! My Tin Tin! (5) [21]

[DEM 1] The Travelator (4.06) [21]

[DAD 1] Electric Boogaloo (3) [19]

[UNC 1] Australian Orienteer (3) [20]

[LIB 1] Carribean Silk (2.12) [20]

[DOT 1] The Amicable Vicar (2.06) [18]

[LIB 3] Green Sakura (2.06) [18]

[CAR 1] Oh Tin Tin! My Tin Tin! elected in position 2 with 5 ballots.

0.06 votes (1 ballot at 0.06 transfer value) originally from [CAR 1] Oh Tin Tin! My Tin Tin! distributed via 4th preference to [UNC 1] Australian Orienteer
0.06 votes (1 ballot at 0.06 transfer value) originally from [CAR 1] Oh Tin Tin! My Tin Tin! distributed via 4th preference to [LIB 1] Carribean Silk
0.06 votes (1 ballot at 0.06 transfer value) originally from [DAD 3] All Aboard distributed via 6th preference [LIB 3] Green Sakura
0.06 votes (1 ballot at 0.06 transfer value) originally from [DOT 6] Prince of Persia distributed via 5th preference to [DAD 1] Electric Boogaloo
0.06 votes (1 ballot at 0.06 transfer value) originally from [LIB 5] Crimson Tide distributed via 5th preference to [DOT 1] The Amicable Vicar

[DEM 1] The Travelator (4.06) [21] {127}

[UNC 1] Australian Orienteer (3.06) [20] {163}

[DAD 1] Electric Boogaloo (3.06) [19] {163}

[LIB 1] Carribean Silk (2.18) [20] {117}

[LIB 3] Green Sakura (2.12) [18] {116}

[DOT 1] The Amicable Vicar (2.12) [18] {156}

Unfortunately for democracy and my insufficient skill in explanation, we've now reached a situation where two ties are tied on 2.12 ballots each, and both measure equally on our first tie-breaker, as they were both preferenced by 18 voters. Thus we turn to our second-tiebreaker measure, the sum of the preferences (as distributed amongst the top 16 candidates), and find that on the whole Green Sakura was a significantly more popular tie amongst the majority of voters than The Amicable Vicar was, garnering a golf-score on 116 compared to a considerably inferior result of 156. (For any of you who object to this reasoning, let it be known that I tossed a coin and it ALSO came up this way, so if prefer your elections vanilla you should also be satisfied that justice was done in this case, and that I also ran a simulation excluding Green Sakura and it ended with the same ultimate result).

[DOT 1] The Amicable Vicar excluded.
1 vote originally from [DOT 1] The Amicable Vicar distributed via 3rd preference to [DEM 1] The Travelator
1 vote originally from [PAI 1] Popular Penguins distributed via 5th preference to [DEM 1] The Travelator
0.06 votes originally from [CON 1] Polytetraflouroethtielene distributed via 6th preference to [DEM 1] The Travelator
0.06 votes originally from [CAR 4] Mouses distributed via 7th preference to [DEM 1] The Travelator

In an impressive dsiplay of solidarity, every vote allocated to the Amiable Vicar went to The Travelator, so at least the voters got more-or-less what they wanted out of this sad exclusion.

[DEM 1] The Travelator (6.18)

[UNC 1] Australian Orienteer (3.06)

[DAD 1] Electric Boogaloo (3.06)

[LIB 1] Carribean Silk (2.18)

[LIB 3] Green Sakura (2.12)

[DEM 1] The Travelator elected in position 3 with 6.18 ballots.

0.23 votes (1 ballot at 0.23 transfer value) originally from [PAI 1] Popular Penguins distributed via 8th preference to [LIB 1] Carribean Silk
0.23 votes (1 ballot at 0.23 transfer value) originally from [DOT 4] Fade to Blue distributed via 4th preference to [LIB 1] Carribean Silk
0.23 votes (1 ballot at 0.23 transfer value) originally from [DEM 1] The Travelator distributed via 5th preference to [LIB 1] Carribean Silk
0.23 votes (1 ballot at 0.23 transfer value) originally from [CAR 1] Oh Tin Tin! My Tin Tin! distributed via 3rd preference to [UNC 1] Australian Orienteer
0.23 votes (1 ballot at 0.23 transfer value) originally from [DEM 1] The Travelator exhausted.
0.23 votes (1 ballot at 0.23 transfer value) originally from [DOT 1] The Amicable Vicar exhausted.
0.014 votes (1 ballot at 0.014 transfer value) originally from [CON 1] Polytetraflouroethtielene via 8th preference [LIB 1] Carribean Silk
0.014 votes (1 ballot at 0.014 transfer value) originally from [CON 1] Polytetraflouroethtielene distributed via 9th preference to [LIB 1] Carribean Silk
0.014 votes (1 ballot at 0.014 transfer value) originally from [LIB 5] Crimson Tide distributed via 8th preference to [UNC 1] Australian Orienteer

[UNC 1] Australian Orienteer (3.304)

[DAD 1] Electric Boogaloo (3.06)

[LIB 1] Carribean Silk (2.894)

[LIB 3] Green Sakura (2.12)

[LIB 3] Green Sakura excluded.
1 vote originally from [LIB 3] Green Sakura distributed via 2nd preference to [UNC 1] Australian Orienteer
1 vote originally from [LIB 3] Green Sakura distributed via 2nd preference to [LIB 1] Carribean Silk
0.06 votes originally from [DAD 3] All Aboard distributed via 14th preference to [LIB 1] Carribean Silk
0.06 votes originally from [CON 1] Polytetraflouroethtielene distributed via 6th preference to [DAD 1] Electric Boogaloo

[UNC 1] Australian Orienteer (4.304)

[LIB 1] Carribean Silk (3.854)

[DAD 1] Electric Boogaloo (3.12)

[UNC 1] Australian Orienteer elected in position 4 with 4.30 ballots.

[LIB 1] Carribean Silk elected in position 5 with 3.85 ballots.

[DAD 1] Electric Boogaloo elected in position 6 with 3.12 ballots.

Since we're dealing with a very small number of ballots here, that plus one in the quota formula, as well as the couple of exhausted ballots, make up a pretty significant proportion of the votes. Thus, the last candidates were elected with significantly fewer votes than the quota, definitely not a mandate, but at least we the people have our representatives. My suitcase will be happily packed this weekend, and I couldn't have done it without you.

Thanks for playing along with our silliness this Blaugust, everybody. I hope you'll stay tuned for Raptember and for whatever this blog turns into when we relocate to England next month.