A common justification for whether or not an act qualifies past the semi-finals is “well, they performed early.” It seems rational: you better remember an act you saw later, and will more likely vote for it. But, prepare for a shock: performance order barely matters.
We at Team Wiwi decided to calculate the linear regression lines of performance orders versus points earned in the semi-finals and finals (in two different equations, obviously). Linear regression lines give us an expected value when we plug in a value into x. All of the data was compiled and given in two equations: semi-final points from 2010 to 2014, and final points from 2010 to 2014 (you try putting in points and running orders for 4 contests in a little calculator!). The equations are as follows:
Semi-Finals: Y= 51.763 + 1.04X (X being the performance order, Y the number of points) Coefficient of Determination: 0.016
Grand Final: Y= 65.787 + 1.866X (X being the performance order, Y the number of points) Coefficient of Determination: 0.04
Hypothetically you could put in the performance order and get a predicted score, but it will not be accurate. The Coefficient of Determination tells how much this line represents the data set (in this case, performance orders and points earned). The closer to 1, the greater the association is, and the closer to 0, the weaker the association. Therefore, there is a very weak association between performance order and points won, and therefore place.
This formula is not perfect. There are countries that place well in every part of the show, due to either a really popular song, suggestive staging, or everyone’s favorite conversation piece: bloc voting. It would be next to impossible to account for all of these variables, so we have to stick to a general formula. Romania’s qualification this year, for instance, would probably not be jeopardized if they performed first in their semi-final. Yet, an entry that is on the cusp of qualifying (cough cough, Slovenia) would have a possibility of benefiting from a later performance.
We now live in an era of Eurovision where quite a few entries can earn 100 points or more, leaving everyone else to fight among themselves. Statistically speaking, though, we cannot pull the “early performance” gimmick. Do you honestly think that Engelbert would’ve placed in the top 10 had he performed last? I think not.
Obviously not everyone agrees with me. David T, a wiwiblogger in the US, has come up with his own analysis where the suggests running order does matter.
What are your thoughts? Is there a relationship between performance order and placing? Tell us in the comments below!
Eurovision and statistics. Such a wonderful combination, and I love you so much for doing it! And kudos for stating the limitations of such an analysis! Have you tried running this with other transformations of the exposure and outcome variables? For example, how about a logistic regression which has an outcome that utilizes 0=did not qualify, and 1=did qualify? Or a rank-order outcome vs. points. Or a transformed exposure variable where you make non-linear assumptions such as as the show goes on, being 2nd vs. 3rd is worse off than being 15th vs. 14th? And then there’s what the other… Read more »
In ant television show the running order is crucial. Yeah some songs would’ve won form pretty much anywhere but others were definetely helped by the running order. It’s not everything but the running order is an important factor in the final placings
Thanks, Davve – A+ for good homework!
I… I got nothing of what you just said Francheska 😀 I mean the whole equation part, but I’ll believe you 😛
No, no one has ever won with start number 2. But I think that is just a coincidence… the songs performed in that number was simply not good enough to win I guess, as both songs from number 1 and 3 has won.
I have a question: performing second always used to be the “graveyard slot” – has anyone ever won ESC performing second in the final? Just curious!
The only bad number you can draw in the running order is the one before or after the winning song.
Agreed with this theory. Malta´s song was not a favorite, but performed first in the semi and still gets through. Or look at Ukraine, who opened this year´s final, still 6th place – a well deserved place, not higher, not lower.
Soooo…. they’re saying Molly still would have come 17th even if she hadn’t performed last even though that some sites said that the UK (obviously some other countries as well) should be disappointed if we didn’t make it to the left hand side of the board? Lies.
I think there’s still a bit of an edge (though not that much) to performing later, at least since televoting started. Looking at the winners alone, 13 of the 17 winners since 1998 (start of televoting) performed in the second half; the 4 winners who performed in the first half were from 1998 (Dana International who got a lot of media attention), 2003 (won by 2 points), 2004 (won by 17 points), and 2014 (Conchita Wurst, who got a lot of media attention). Still, a lot of the bad songs that placed last did so because they simply sucked. Suppose… Read more »
After almost 60 years of Eurovision (yes kids … 60 … not 16 …) it’s unbelievable how people still believe in craps of this nature … voting (especially the televoting) couldn’t care less if you are number 1 or 14, if it’s better or worse, the placing is not irrelevant … what matters is what number you are so that they can cast a vote and that it’s, and in that case being a 7 or a 20 means the same thing. But if that is still earth shattering traumatizing apocalyptic depressive factor in your life … do humanity a… Read more »