![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
danieldwilliamI’m going to take a look at the checks and balances within the Strictly Come Dancing Electoral College. The purpose of the College is to select the couple who will leave the competition. I think there is a slight systemic bias in favour of weaker couple. This is probably reduced in practise by a correlation between the popular vote and the quality of the performance.
There are four elements to the Strictly Come Dancing Electoral College. The judges’ scoring, the public vote, the combination of the these two elements and the dance off.
The judges’ scoring is very straightforward. Four judges award marks from 1-10 based on whatever criteria they like. The scores are amalgamated to give each couple a score from 4-40. There is no special role for any particular judge. A point is worth a point regardless of how it is given or by whom. Four 9’s are worth the same as two 8’s and two 10’s.
The judges’ scores inform a leader board, top scoring couple at the top, lowest scoring couple at the bottom. Points are awarded based on this ranking. The highest scoring couple scores points equal to the number of couples remaining in the competition. So if there were 8 couples left the top couple would score 8 points. The second place couple would score 7 and the lowest place couple would score 1. Couples who are tied, tie up, i.e. the two couples are awarded jointly the best place. Two couples on the same score, 1 point behind the leader get joint second, rather than joint third. Everyone else moves up a place. This tends to favour weaker dancers. There is a good chance that two or more couple will end up on the same score. This pulls the bottom placed couple up and they end up with more than 1 point. We’ll see how this works to their advantage when we turn to the public vote.
Four judges scoring independently provide a bit of balance to each other. The judges’ scoring also give the public votes something to mill on.
The second stage of the Electoral College is the public vote. Couples are ranked in order of their public vote on the night. This ranking is converted into points in the same way as the ranking from the judges’ scores are converted into points. Given the very small possibility that couples will be tied on votes the rules on tie-ing up don’t favour unpopular couples in the way they favour poorer dancers.
This gives us two leader boards both converted into points. To create the final leader board the points are added together to create a Combined Leader Board. This balances up the professional opinions of the judges with the popular view of the public at large. (Sort of, I’d prefer that we used preferential voting for each round of the public vote.) In practise it is difficult for someone who finishes top to end up in the dance off. It’s actually fairly easy for the bottom placed couple to avoid the dance off.
The bottom two couples enter the dance off. There are no tied places here. In the event of a tie the couple with the largest public vote ranks higher. This is where the rules on tie-ing up favour weaker dancers too.
Firstly, a couple who finish bottom of the judges scoring but top the popular vote will finish ahead of a couple who top the judges scoring but finish last in the public vote. In a week with 8 couples the top ranking dancer would score 8 from the judges and 1 from the public vote. The bottom ranked dancers will pick up 1 point from the judges leader board and 8 points from the public vote. Both couple score 9 in total. The weaker, more popular dancers ranks ahead based on the public vote. In situations where the top ranked for dancing couples are close on Combined Points with the weak dancers the stronger dancers must necessarily have performed worse in the public vote.
The second way the College favours weaker dancers is that they are more likely to be advantaged by the way tied couples tie-up. In a week with 8 couple and two ties, say for 2nd and 4th spots, then the bottom ranked dancer takes 3 points into the Combined Leader Board from the dancing. There is no way the top ranked dancer can score more than the maximum points, but the bottom ranked dancer could score more than the minimum points. Imagine a tie between the top two couples and the bottom two couples in an 8 couple week.. In this case if the best two dancing couples finished bottom and second bottom of the popular vote and the bottom two topped the voting they would end up with a combined scores as follows. Top Dancers, Lowest Votes, 8 +1 = 9, Top Dancer Second Lowest Vote, 8 + 2 = 10. Lowest Dancers, Top Vote, 3 + 8 = 11 points, Lowest Dancer, Second Top Vote, 3 + 7 = 10 votes. Both of the top dancers go into the dance off.
If I were reforming the Strictly Electoral College I would change the way the judges’ leader board deals with ties by creating a tie-breaker. I’d probably start with highest numbers of 10’s, then 9’s in this week’s scoring then cumulatively. So a couple with two 8’s and Two 10’s beats a couple with four 9’s. Starting with the cumulative number of 10’s might also work as it favours couples with consistently good scoring.
The two bottom ranked couples go in to the last stage of the College which is the dance off. They perform again. The judges vote on who to save. There are in effect five votes. Len as head judge has two. This will tend to favour stronger dancers. The judges have the final filter and will, by definition save the best couple of the two in the dance off. Whether they save the best couple on the night or the best couple over all is a matter for their conscience. However, they can only save the best of the two couples offered up to them. As we’ve seen with John Sergeant and Anne Widdicombe popular but poor dancers can avoid the dance off all together for weeks.
The judges can signal their disdain by awarding four 1’s to a dancer but his still puts them in bottom place and gives them 1 or more Electoral College votes in the Combined Leader Board.
I think the rules of the Electoral College slightly favour weaker dancers particularly where they have strong public backing not connected to the quality of their dancing. This does make the competition more of a popularity contest than perhaps it first appears. In practise this is probably mediated by the general tendency of the public to vote for the better dancers and the fact that over the long term the competition is really a run off voting system. A weak but popular dancer will eventually succumb to the Electoral College as their progress through the rounds begins to threaten better dancers and the vote for better dancers stiffens.
But then John Sergeant did happen.
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
no subject
Date: 2012-12-03 12:44 pm (UTC)no subject
Date: 2012-12-03 01:13 pm (UTC)I wouldn’t want it to become more of a popularity contest for example by removal of the dance off.
no subject
Date: 2012-12-03 06:19 pm (UTC)no subject
Date: 2012-12-03 01:00 pm (UTC)This is probably reduced in practise by a correlation between the popular vote and the quality of the performance.
I am fascinated by this assumption. I would have assumed the precise reverse, from a combination of two foundational assumptions: firstly, that popular attachment to celebrities is entirely a function of their looks, charm and possibly claim to fame and that this and dancing ability are independent variables; and secondly that poor dances and low scores activated voters who would not bother to vote if their chosen celebrity had done better in the judge's rankings.
no subject
Date: 2012-12-03 01:43 pm (UTC)I think it might be safe to say that given two celebrities of equal popularity the one with the most dancing talent is likely to pick up a few extra votes. I know to my cost how important a few votes here or there can be.
I think there are two elements to a celebrity’s popularity as it translates into Strictly votes. There is the initial popularity and then developing dance popularity.
I’m going to treat initial popularity as static and lump both the public’s reaction to actual dancing talent and any additional insight into character the public see brought out by The Dance as developing. Just so I can get my head round it. You PhD’s feel free to think in 3D.
Put it another way, initial popularity can be damaged or enhanced by floor performance and the narrative arc of the JuurNAY of the celeb.
(There is also a third factor which is propensity of the demographic who like a particular celebrity to turn out for them. I don’t think people who remembered Johnny Ball were that likely to turn out for him. I suspect Denise’s fan base might not be a high turn out demographic either.)
So, people like soap actors or well liked national treasures will tend to do well. But every week the folk see more and more character in both senses and they also, I suggest, begin to base their voting more on a fair assessment of merit. I’d hold Chris Hollins up as an example of this. Middling celebrity before SCD. Did very well, and surprisingly so with his Charlston (Ola’s Charleston really) and that, I think, gave him the developing popularity to win through.
So what I think is going on is that celebrities (and pro-dancers) arrive with an endowment of popularity. This is going to be a function of all the things you mention. To this is added (or subtracted) a change in popularity. I think this is influenced in part by the actual dancing abilty. (I hold myself up as an example. Denise is my favourite celebrity this year regardless of whether she is any good. However, if she were in danger I would turn out to save her because if someone who tops the scoring most weeks doesn’t make it to the final that is wrong. I’m struggling to think of many people I would vote to keep in even if they were really poor.)
I definitely concur that doing badly in the judges’ scoring is likely to motivate voters to turn out. I think this might be a termporary phenomenon. One bad week gets rescued by popular support. Four in a row less so.
But again John Sergeant. (But I could suggest that he is a special case of character – inter alia – being revealed through The Dance.)
Difficult to draw any conclusions with out the voting dating. I’m seriously considering an FOI request to the BBC.
no subject
Date: 2012-12-03 01:57 pm (UTC)This is the bit I disagree with, I think. Kimberley, always a fab dancer, wasn't getting votes until something bad happened to her and people saw a more vulnerable side to her character. I don't think Victoria lost votes for not being a good dancer, but for being a whiner. Similarly I think Michael gained votes because he was seen to be working his guts out, rather than because he got all that much better. But all this is my interpretation only and could be wrong.
I didn't see the Sergeant series so can't comment on that.
It's certainly true by the time of the two semis and finals I've seen, the people who lasted through popular vote or character (Gavin, Pamela, Alex, perhaps Chelsee) were voted down in favour of the best dancers (Kara and Harry). So if that trend lasts, it'll be Denise or Dani.
no subject
Date: 2012-12-03 02:38 pm (UTC)(Hence the need for my FOI request.)
Or Michael. (Who I feel a bit sorry for, if he’d survived and then gotten the two ballroom dances in a row he would be owed I think he could have snuck through to the semi’s.)
Sergeant was legendary. He was a very bad dancer but was having such a great time doing it and losing such a lot of weight and so pleased about that the public had some sympathy for him. His big moment was a paso doble in week 6 where Kristina Rihanoff (in her first series) pulled off some blinding choreography including being dragged across the dance floor by a strutting John Seargeant. She followed this up week after week with some imaginative choreography and basically dancing round JS at about twice the speed of sound. He pulled out when it looked like he might win.
no subject
Date: 2012-12-03 06:25 pm (UTC)no subject
Date: 2012-12-03 08:03 pm (UTC)no subject
Date: 2012-12-03 06:24 pm (UTC)no subject
Date: 2012-12-03 08:03 pm (UTC)no subject
Date: 2012-12-03 01:43 pm (UTC)I endevour to give satisfaction, ma'am.
no subject
Date: 2012-12-03 06:06 pm (UTC)no subject
Date: 2012-12-03 08:05 pm (UTC)no subject
Date: 2012-12-03 08:05 pm (UTC)no subject
Date: 2012-12-03 08:35 pm (UTC)But in some ways it seems simple:
Judges vote, points allocated in order of score.
Public votes, points allocated in order of score.
Score of contestant = Judge Points + Public Points.
I think that giving the tie-break to the public gives people more of an incentive to vote - and that handing the tie-break to the score given by the elitist ruling classes is a disincentive to direct democracy :->
no subject
Date: 2012-12-04 09:55 am (UTC)It does seem simple. It’s made up of a small number of pretty straightforward elements. All of the elements seem intuitively fair and simple. Judges judge, public votes, add ‘em, together, bish bash bosh. I don’t think at any point does it go non-linear.
Not sure if you’re refering to the tie-break in the Dance Off or the way tie-breaks in the judges’ scoring is handled.
I have been thinking about direct democracy quite a lot recently and if I have time will put up some of my posts on the subject. I think we should have more direct democracy in the UK. However, I quite like the idea of liquid democracy and there is an element of liquid democracy in the use of the judges to pick from the bottom two.
My proposed “improvement” for the handling of tie-breaks in the judges’ scoring is aimed at reducing the slight advantage that lower placed couples get by forcing a pure ranking rather than having tied places. It’s only an improvement if you feel that favouring weaker dancers is a bug, not a feature of the system. Reasonable people could disagree about the extent to which Strictly is a dancing programme with some celebrities or a celebrity programme with some dancing.