The EBU has recently announced a new change in the voting system for 2018, with the way the jury voting is handled. From linear to ‘exponential’ ranking. Not sure what that means? We’ll explain.
First, how did the jury work before?
Since 2013, the jury voting was fairly simple. Let’s take a very specific example: the French jury in the 2016 final. Five jurors ranked all the songs of the night (25, since they couldn’t vote for France). And then the five individual rankings were made into an average Top 25.
An “average jury top 25” is made by calculating the average rank of each song for the jury. For example, in 2016, the French jurors ranked Spain 8th, 5th, 7th, 19th, and 2nd. On average, that makes “8.2”. You just calculate that for the 25 songs, and then you rank them from best to worst average rank. Here, 8.2 was the 8th best average rank (behind Malta’s 8 and ahead of Armenia’s 11.8), so Spain was 8th in the jury ranking, and thus earned 3 points from France.
In an even easiest way to understand, you can also think of it as “adding the ranks” (for Spain, 8+5+7+19+2=41), and then the top 10 consists of the songs with the lowest scores (obviously, if you’re the favorite of all five jurors, the added ranks make 5 (1+1+1+1+1), so you’re pretty low)
Well, now, say bye to the average. The jury will be going exponential! The actual technique used is called “exponential smoothing”. So, what does that mean? It means that each rank is given a “score” (a weight), which is higher with a high rank, and lower with a low rank. Let’s call those “juror scores“. So the five jurors will still make a Top 25/26, and for each individual ranking, the songs will score depending on their rank. You then just add the five juror scores of each song, and you have a common Top 25/26!
This graph shows approximately the scores each rank is given. We don’t know yet which weight is given to each rank, but we know that the favorite of a juror will score 12 points. And, thanks to some good approximations made with the graph, we can estimate those weights:
|Juror rank ||Juror Score |
This is a double approximation: the weights are approximated using a mathematical formula which make it as close as possible to what we see in the graphs, and to make it easier to read, they have been simplified to the 2nd decimal (0.42 is easier to read than 0.4215208).
Why does it matter?
It matters because this system rewards any song that a juror like, and attenuates the influence of a low ranking.
As you can see in the estimation, the gap in score between each rank is shorter when the rank is low. For example, between 1st and 2nd, the gap is of 1.5 juror points. But between the 11th and the 12th, it’s only 0.4 (from 3.09 to 2.69), and between 21st and 22nd, only 0.1. So the higher you rank, the even higher you score. And wether you’re 21st or 14th doesn’t change your score a lot.
Basically, if a song is ranked very high by one or two jurors, and very low by the others, it will still get a very decent juror score, and thus possibly be in the Top 10 of the jury as a whole.
Not convinced? Let’s take a real example
Let’s take our French jury for Stockholm 2016, and their ranking of Ukraine. From best to worst: 2nd, 10th, 12th, 21st, 24th. One juror had Ukraine second best, another had Ukraine 2nd least favorite (France being in the final, the jurors gave a Top 25).
At the time, this gave Ukraine a 15th place in the French jury ranking. But with the new system, the 2nd place of a juror has more weight than the 24th of another: as a whole, Ukraine would have scored 18.08 juror points, which would actually place it 9th, and thus earn 2 points from the French jury.
The same thing would have happened to Israel (from best to worst: 2nd, 9th, 21st, 22nd, 23rd), which was 17th in the jury ranking, but with a score of 16.68, would be 10th with the new system, and earn 1 point from the French jury.
Here is the full impact on this jury:
So, which songs are more impacted?
Divisive songs. A song that strongly divides a jury will have more chance to get in the top 10 it it creates a strong positive feeling for at least two jurors (Israel for France in 2016, for example), since the three others will have less power to rank it down.
A song that creates strong positive consensus will still gain more points than a divisive song. However, a song around 10th by all five jurors could more easily miss out the Top 10. Songs that the jurors could see as “allright”, or “safe”. In the French example, this is the case of the Czech Republic, which was already 11th (in details: 11th, 12th, 12th, 14th, 17th), but would fall to the 17th place with the new system.
Hence why this new system rewards risk, and why it could entice delegations and broadcasters to internally select risky acts.
Basically, the new system might not change anything for “masterpieces” or “absolute last songs”, but gives an advantage to “risky songs” over “safe songs”.
What do you think of this new system? Do you find it fairer? What could be the risks, or the hidden advantages? Tell us more on the comments below or on social media at @ESCXTRA !